Integrating multitrait physiological approaches to characterize salinity tolerance in bread wheat (Triticum aestivum L.) genotypes

Md Mustafa Khan; Md Abdul Karim; Md Moynul Haque; M A Baset Mia; Naresh Chandra Deb Barma; Hela Znazen; Ahmed Gaber; Akbar Hossain

doi:10.1186/s12870-026-09756-4

. 2026 Aug 17;26:1639. doi: 10.1186/s12870-026-09756-4

Integrating multitrait physiological approaches to characterize salinity tolerance in bread wheat (Triticum aestivum L.) genotypes

Md Mustafa Khan ^1,^✉, Md Abdul Karim ^2,³, Md Moynul Haque ², M A Baset Mia ⁴, Naresh Chandra Deb Barma ^1,⁵, Hela Znazen ⁶, Ahmed Gaber ⁷, Akbar Hossain ^8,^✉

PMCID: PMC13613665 PMID: 42786434

Abstract

Background

Soil salinity is a major constraint to wheat production because it disrupts plant water relations, photosynthesis, oxidative balance and yield formation. Although numerous physiological traits have been investigated individually, integrative approaches for evaluating coordinated physiological responses to salinity remain limited. This study investigated whether multitrait physiological integration could contribute to the characterization of salinity tolerance in the evaluated wheat genotypes using conventional physiological analyses together with cost–benefit slope analysis and the integrated physiological efficiency index (IPE_ridge).

Results

Three contrasting bread wheat genotypes [BAW 1147 (tolerant), BARI Gom 25 (moderately tolerant) and BARI Gom 28 (susceptible)] were evaluated under control, 5 and 10 dS m^-1 electrical conductivity (EC) of the irrigation solution in a split-plot experiment arranged in a randomized complete block design. Water relations, gas exchange, photosynthetic pigments, compatible solute accumulation, oxidative stress, antioxidant defence and yield-related traits were assessed and physiological coordination was further evaluated using cost–benefit slope analysis and the integrated physiological efficiency index (IPE_ridge). Salinity (EC) of the irrigation solution reduced plant water status, photosynthetic performance, pigment stability and yield while increasing compatible solute accumulation, oxidative stress indicators and antioxidant enzyme activities. Among the evaluated genotypes, BAW 1147 generally maintained higher relative water content, photosynthetic rate, water use efficiency, chlorophyll content and grain yield, together with lower oxidative damage, than BARI Gom 25 and BARI Gom 28. Correlation analysis suggested stronger coordination among physiological and yield-related traits under saline irrigation conditions than under control conditions. Cost–benefit slope analysis revealed genotype- and irrigation solution EC-dependent differences in the relative contributions of the physiological response modules, whereas IPE_ridge values were generally higher in the tolerant genotype than in the moderately tolerant and susceptible genotypes.

Conclusions

The findings suggest that salinity tolerance in the evaluated wheat genotypes was associated with coordinated physiological responses rather than the performance of individual traits alone. The proposed cost–benefit slope analysis and IPE_ridge index may provide complementary quantitative approaches for characterizing multitrait physiological responses within the present dataset. However, further validation across larger and more diverse wheat populations and environmental conditions is required before their broader applicability can be established.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12870-026-09756-4.

Keywords: Salinity stress, Wheat, Osmotic adjustment, Oxidative stress, Antioxidant enzymes, Water use efficiency, Photosynthetic traits, Multitrait coordination, Cost–benefit analysis, Integrated physiological efficiency (IPE_ridge)

Introduction

Soil salinity is among the most pervasive abiotic stresses limiting global wheat production; more than 830 million ha of land worldwide is affected by salinity and the extent of salt-affected areas is projected to increase because of climate change-driven sea-level rise, secondary salinization and irrigation mismanagement [35, 41]. In coastal South Asia, where saline groundwater and tidal intrusion increase soil electrical conductivity (EC), the development and cultivation of salt-tolerant wheat cultivars are important for supporting food security and the sustainable use of affected land [42]. Wheat (Triticum aestivum L.) exhibits considerable genotypic variation in terms of its physiological responses associated with the osmotic, ionic and oxidative components of salinity stress [15, 34].

During the initial osmotic phase, elevated external NaCl concentrations lower soil water potential and impose a physiological water deficit that restricts water uptake, reduces leaf water potential (LWP) and limits turgor-dependent cell expansion [24, 54]. Salinity-induced osmotic stress also activates signalling pathways involved in water transport and compatible solute biosynthesis, thereby contributing to plant responses under low-water-potential conditions [56]. The maintenance of relatively high relative water content (RWC) and less negative LWP under salinity may indicate an improved capacity to preserve plant water status. This response can be associated with the accumulation of compatible solutes, particularly proline and soluble sugars, which contribute to cellular osmotic balance while stabilizing proteins and membranes and interacting with redox metabolism to protect cellular functions under stress [3, 9, 26, 27, 50, 51, 59]. Previous studies have reported that compared with susceptible wheat genotypes, salt-tolerant wheat genotypes often maintain better water status and accumulate greater concentrations of proline and soluble sugars [20, 21].

During prolonged exposure, salinity may also cause ionic imbalance through the excessive accumulation of Na^⁺ and Cl^-, thereby disrupting cellular homeostasis and metabolic processes [35, 44]. In wheat, salinity tolerance is frequently associated with the regulation of ion transport and the maintenance of favourable K^⁺:Na^⁺ homeostasis [23, 36]. However, ion concentrations were not measured in the present study. Therefore, the current investigation focuses on physiological and biochemical responses related to plant water status, compatible solute accumulation, photosynthetic performance, oxidative stress and antioxidant defence.

Salinity can also induce oxidative stress through the excessive production of reactive oxygen species (ROS), including hydrogen peroxide (H₂O₂) and superoxide radicals, which may damage lipids, proteins, membranes and the photosynthetic apparatus [18, 33]. At controlled concentrations, ROS also function as signalling molecules that activate antioxidant defence pathways and other stress-responsive processes [43, 47]. The extent of oxidative injury varies among genotypes and is influenced by antioxidant defence systems that include enzymatic components, such as catalase (CAT), ascorbate peroxidase (APX) and glutathione peroxidase (GPX), together with nonenzymatic antioxidants such as carotenoids [33]. Greater antioxidant activity may help maintain redox balance, protect photosynthetic pigments and sustain photosynthetic functioning under saline conditions [12, 39]. Previous studies have shown that compared with susceptible genotypes, salt-tolerant wheat genotypes generally exhibit greater antioxidant enzyme activity, lower oxidative damage and increased compatible solute accumulation, suggesting that these responses are coordinated with salinity adaptation [20, 21]. Enhanced antioxidant capacity, including increased CAT and APX activities, has also been associated with reduced oxidative injury and improved stress tolerance in crop plants exposed to salinity [25].

The combined effects of these physiological and biochemical responses ultimately influence plant performance under saline conditions. The grain yield represents the cumulative outcome of the plant water status, photosynthetic capacity, compatible solute accumulation, oxidative damage and antioxidant protection rather than the contribution of a single trait [6, 52]. Accordingly, comparisons among genotypes may benefit from analytical approaches that evaluate the coordination of multiple physiological responses associated with photosynthetic performance and yield under salinity stress.

Traditional analyses typically evaluate physiological traits individually or examine their relationships using correlation, principal component analysis, clustering, or other multivariate approaches. More recently, integrative frameworks combining morpho-physiological, biochemical and multivariate analyses have highlighted the value of integrating multiple complementary indicators for characterizing salinity responses across different crop species [16, 17]. Although these studies were conducted in crops other than wheat, they illustrate the potential of integrated physiological indicators and multivariate analyses for improving stress characterization and genotype evaluation. Building on this concept, complementary analytical descriptors may provide additional insight into how coordinated physiological responses are associated with functional outcomes within a specific dataset. In the present study, cost–benefit slope analysis was used to examine changes in selected physiological response modules relative to the photosynthetic rate and grain yield across salinity transitions. In addition, a ridge regression-based integrated physiological efficiency index (IPE_ridge) was used to combine correlated physiological variables while reducing the instability associated with multicollinearity [11, 19]. These approaches were applied as exploratory and complementary tools for characterizing multitrait physiological coordination in the evaluated genotypes rather than as universally applicable selection methods. Their broader applicability, however, requires independent validation using larger and more diverse plant populations and environmental conditions.

Accordingly, this study aimed to characterize the physiological responses associated with salinity tolerance in three contrasting wheat genotypes-BAW 1147 (tolerant), BARI Gom 25 (moderately tolerant) and BARI Gom 28 (susceptible)-under control conditions and salinity (EC) of 5 and 10 dS m^-1 of the irrigation solution. The specific objectives were to (i) characterize genotype-specific changes in plant water relations, compatible solute accumulation, oxidative stress and antioxidant defence under salinity; (ii) examine the associations of these physiological responses with water use, photosynthetic performance and grain yield and (iii) evaluate multitrait physiological coordination in the evaluated genotypes using cost–benefit slope analysis and the IPE_ridge framework.

Materials and methods

Experimental site

The pot experiment was conducted from November 2013 to March 2014 under an open-sided, polyethylene-covered lath house at the Agronomy Research Field of Bangabandhu Sheikh Mujibur Rahman Agricultural University (BSMRAU) (currently Gazipur Agricultural University), Gazipur, Bangladesh. The structure excluded direct rainfall while allowing natural air circulation through the open sides, thereby minimizing heat accumulation and maintaining environmental conditions comparable to the ambient atmosphere. The site is located at approximately 24°02′ N latitude and 90°23′ E longitude, with an elevation of approximately 12 m above sea level. The region is characterized by a subtropical climate with distinct seasonal variations. The winter growing period (November–March) is typically cool and dry, with low relative humidity and moderate temperatures, providing suitable conditions for wheat cultivation under controlled irrigation. During the experimental period, the maximum, minimum and mean air temperatures were 28.2 °C, 12.7 °C and 20.5 °C, respectively.

Plant materials and salinity treatment application

Three hexaploid bread wheat (Triticum aestivum L.) genotypes with contrasting salinity tolerance were selected: BAW 1147 (tolerant), BARI Gom 25 (moderately tolerant) and BARI Gom 28 (susceptible), following the classification of Khan et al. [28]. Seeds were obtained from the Regional Station, Bangladesh Wheat and Maize Research Institute (BWMRI), Gazipur, Bangladesh. This station was formerly known as the Regional Wheat Research Centre (RWRC) under the Bangladesh Agricultural Research Institute (BARI) before the establishment of the BWMRI.

The salinity treatments consisted of three irrigation solution EC levels: control (tap water; EC = 0.35–0.40 dS m^-1), 5 dS m^-1 (moderate salinity) and 10 dS m^-1 (high salinity). The control water had an approximately neutral pH (≈ 7.0). Natural seawater with an EC of approximately 32 dS m^-1, collected from the Bay of Bengal, was used as the salt source and diluted with tap water to prepare irrigation solutions with target EC values of 5 and 10 dS m^-1. The EC of the control and saline irrigation solutions was verified using a calibrated EC meter (HI98304; Hanna Instruments, Romania). The ionic composition of the original seawater was characterized before dilution. The predominant cations were Na⁺ (2840 ppm), Ca2⁺ (542 ppm), Mg2⁺ (526 ppm) and K⁺ (213 ppm). Their concentrations were determined using an atomic absorption flame spectrophotometer (Model AA240FS; Agilent Technologies, Australia) at the Soil Science Division Laboratory, Bangladesh Agricultural Research Institute, Gazipur, Bangladesh.

Salinity treatments were initiated at 10 DAS by gradually applying the designated saline solutions to minimize osmotic shock. Thereafter, the pots were irrigated throughout the growing period with their respective treatment solutions according to the crop water requirements. At each irrigation event, all the pots, including the control, received an equal volume (2 L pot^-1) of the designated irrigation solution. The pots had drainage holes to prevent waterlogging and allow excess water to drain. The EC values reported for the treatments refer to the applied irrigation solutions, which were maintained at the target EC levels throughout the experiment. However, the root-zone soil EC, leachate EC, drainage volume and leaching fraction were not monitored during the experimental period.

Experimental design and seed sowing

The experiments were arranged in a split-plot design within a randomized complete block design (RCBD) with three replications (n = 3). Saline irrigation solutions including control were assigned to the main plots, while genotypes were allocated to the subplots, resulting in nine treatment combinations. Uniform, healthy and well-developed seeds of each genotype were manually sown on 25 November 2013 in round plastic pots (30 cm diameter × 30 cm depth) placed within the lath house. Each pot was filled with 16 kg of sandy-loam soil mixed with well-decomposed, air-dried cow dung at a 3:1 ratio (soil:cowdung). The growing medium was thoroughly homogenized and premoistened prior to pot filling. Eight seeds were sown per pot, followed by light irrigation to ensure uniform germination. At 10 days after sowing (DAS), the plants were thinned to two uniform plants per pot.

Fertilizer application

Fertilizers were applied at recommended field rates of 100–60-40–20-1 kg ha^-1 for N-P₂O₅-K₂O-S-B. On a per-pot basis, 2.0 g of urea, 1.25 g of triple superphosphate (TSP), 1.0 g of muriate of potash (MoP), 1.0 g of gypsum and 0.06 g of boric acid were used. At sowing, two-thirds of the urea and the full amounts of the remaining fertilizers were thoroughly incorporated into the soil. The remaining one-third of the urea was applied as a top dressing at the crown root initiation (CRI) stage (approximately 18 DAS), followed by irrigation to facilitate nutrient availability and uptake.

Intercultural operations

All the pots were managed uniformly throughout the experimental period. Intercultural practices included manual weeding and standard plant protection measures as needed. To minimize positional effects within the lath house, the pots were periodically rearranged to ensure uniform exposure to environmental factors such as light, temperature and airflow.

Plant water relations

Measurement of leaf water potential

The leaf water potential (LWP; MPa) was measured before sunrise to minimize transpirational gradients and diurnal variability. Measurements were conducted under predawn field conditions when transpiration was minimal and the leaf water status was close to equilibrium with the soil. At anthesis and grain filling, the fully expanded flag leaf (one plant per genotype × treatment × replicate) was excised at the base with a new razor blade and inserted into a Scholander-type pressure chamber (compressed N₂) within ~ 10–20 s of cutting to limit evaporative artifacts [45].

The leaf base was gently protruded through the rubber gasket (~ 1–2 mm exposed) to ensure an airtight seal without tissue damage. The chamber pressure was increased at a controlled ramp rate (≈0.02–0.05 MPa s^-1), while the cut surface was viewed under a hand lens. The endpoint was the first appearance of xylem sap (meniscus) at the cut surface; pressure was then immediately released and the balancing pressure (P_bal) was recorded. LWP was reported as Ψleaf = − P_bal (MPa). All samples and the chamber were shaded during handling to avoid heating; instrument zero, gasket integrity and leaks were checked prior to measurements.

Measurement of relative water content

Relative water content (RWC%) was determined at anthesis and grain filling from the fully expanded flag leaf (one plant per genotype × treatment × replicate), sampled predawn and sealed in zipper bags on ice for rapid transfer to the laboratory. Fresh weight (FW) was recorded immediately after arrival. The leaves were then rehydrated for 24 h in deionized water at room temperature under low light, blotted gently to obtain the turgid weight (TW) and oven-dried to a constant mass at 70 °C for 72 h to obtain the dry weight (DW). The RWC was calculated as follows: Inline graphic [46].

Compatible solute estimation

Estimation of proline content

Analyses were performed at the Plant Breeding Division, Bangladesh Agricultural Research Institute (BARI), Gazipur. At anthesis, fully expanded flag leaves (one per genotype × treatment × replicate) were harvested predawn, sealed on ice and brought to the laboratory. Approximately 0.5 g of the upper lamina was homogenized in 10 mL of ice-cold 3% (w/v) sulfosalicylic acid and centrifuged; the supernatant was subsequently used for the assay. Following Bates et al. [4], equal volumes (1:1:1) of extract, acid ninhydrin reagent (ninhydrin in glacial acetic acid plus mineral acid) and glacial acetic acid were incubated at 100 °C for 60 min and then cooled on ice. The chromophore was extracted with 4 mL of toluene and the absorbance was read at 520 nm (UV-1800, Shimadzu). The proline concentration was determined from an L-proline standard curve and is expressed as μmol g^-1 FW using a molecular mass of 115.5 g mol^-1.

Estimation of soluble sugars

Analyses were performed at the Post Harvest Technology Division, BARI, Gazipur. Flag leaf samples collected at anthesis were oven-dried at 70 °C for 72 h and then finely ground. 0.05 g (50 mg) powder was extracted in 80% (v/v) ethanol (multiple washes, pooled to a known volume). For colorimetry, an aliquot of extract was mixed with fresh anthrone reagent (0.2% anthrone in concentrated H₂SO₄), heated in a boiling-water bath for 7.5 min and then cooled to 25 °C. The absorbance was recorded at 630 nm using a UV–Vis spectrophotometer (T80, PG Instruments, UK). The soluble sugar concentration was calculated from a glucose standard curve and expressed as mg g^-1 dry weight [57].

Gas exchange measurements

Gas exchange was assessed at anthesis on fully expanded flag leaves of each genotype across all salinity treatments, with measurements taken on all replicates. The experimental pots were carefully transported from the experimental lath house to the Bangladesh Rice Research Institute (BRRI), Gazipur, using an office pick-up vehicle and gas exchange measurements were performed outdoors under clear-sky conditions between 11:00 and 13:00 (local time). During the measurement period, the photosynthetically active radiation (PAR) at the leaf surface ranged from 1100 to 1200 µmol m^-2 s^-1. An LI-6400 portable photosynthesis system (LI-COR Biosciences, Lincoln, NE, USA), owned by the Physiology Division of BRRI, was used under ambient environmental conditions with an ambient CO₂ concentration of approximately 360 µmol mol^-1 to determine the net photosynthetic rate (P_n; µmol CO₂ m^-2 s^-1), stomatal conductance to water vapour (g_sw; mmol H₂O m^-2 s^-1), intercellular CO₂ concentration (C_i; µmol mol^-1) and transpiration rate (E; mmol H₂O m^-2 s^-1). Stomatal conductance to CO₂ (g_sc; mmol CO₂ m^-2 s^-1) was calculated from the machine-reported g_sw using the standard diffusivity ratio (g_sc = g_sw/1.6). Leaf-level water use efficiency (WUE) was calculated as the ratio of P_n to E for each observation and replicate-level data were retained for statistical analyses.

Photosynthetic pigment determination

The chlorophyll and carotenoids contents were determined at the anthesis stage using fully expanded flag leaves of each genotype across all treatments. For each genotype, three flag leaves were randomly sampled. Immediately after excision at the base, the leaf blades were placed in zipper-sealed polyethylene bags, kept in an ice box and promptly transferred to the Molecular Laboratory of the Plant Breeding Division, BARI, Gazipur.

Leaf pigments were extracted following the methods of Arnon [2] with slight modifications. Briefly, 0.5 g of fresh leaf tissue from the upper portion of the flag leaf was homogenized in 10 mL of 80% (v/v) acetone under chilled conditions (4 °C). The homogenate was centrifuged at 2500 rpm for 10 min at 4 °C and the extraction was repeated until the residue became colourless. The combined extracts were adjusted to a final volume of 10 mL with 80% acetone. Pigment quantification was performed immediately using a UV–Visible spectrophotometer (Model UV-1800, Shimadzu, Japan). The absorbance values were recorded at 663, 645 and 480 nm against 80% acetone as the blank.

Chlorophyll a (Chl a), chlorophyll b (Chl b), total chlorophyll and carotenoids were calculated according to the equations of Arnon [2] and Lichtenthaler and Wellburn [31]:

Chlorophyll a (μg mL^-1) = 12.7 × A₆₆₃ − 2.69 × A₆₄₅
Chlorophyll b (μg mL^-1) = 22.9 × A₆₄₅ − 4.68 × A₆₆₃
Total chlorophyll (μg mL^-1) = 20.2 × A₆₄₅ + 8.02 × A₆₆₃
Carotenoids (μg mL^-1) = (1000 × A₄₈₀ − 1.82 × Chl a − 85.02 × Chl b)/198

where A₆₆₃, A₆₄₅ and A₄₈₀ denote the absorbance at 663, 645 and 480 nm, respectively and Chl a and Chl b in the equation of carotenoids represent the chlorophyll concentrations calculated from the corresponding equations. The pigment concentrations obtained from the above equations were converted to a fresh-weight basis using the following equation:

where C is the pigment concentration in the extract (μg mL^-1), V is the final extract volume (10 mL) and W is the fresh weight of the extracted leaf tissue (0.5 g) and a microgram-to-milligram conversion factor (1000). The final chlorophyll a, chlorophyll b, total chlorophyll and carotenoids contents are expressed as mg g^-1 fresh weight (FW).

Oxidative stress indicators

Estimation of H₂O₂

The hydrogen peroxide (H₂O₂) content was determined following the methods of Yu et al. [58] with slight modifications. Fresh leaf tissue (0.5 g) was homogenized in 3 mL of 50 mM potassium phosphate buffer (pH 6.5) under chilled conditions (4 °C). The homogenate was centrifuged at 11,500 × g for 15 min and the resulting supernatant (3 mL) was collected. Subsequently, 1 mL of 0.1% titanium chloride (TiCl₄) prepared in 20% (v/v) H₂SO₄ was added and the mixture was centrifuged again at 11,500 × g for 15 min at room temperature. The absorbance of the supernatant was measured spectrophotometrically at 410 nm.

The H₂O₂ concentration was calculated using an extinction coefficient of 0.28 μM^-1 cm^-1 and expressed as µmol g^-1 FW.

Estimation of cell membrane injury

Cell membrane stability was estimated following the procedure of Blum and Ebercon [7] with minor modifications. Fully expanded flag leaves were sampled at noon from three replications of all treatment combinations by randomly collecting three leaves from two plants per replication. The samples were thoroughly washed three times with deionized water to remove surface contaminants and gently blotted dry with tissue paper.

From each bulked sample, two sets of ten leaf discs (10 mm diameter each) were excised using a leaf punch. Each set of discs was placed in a 20-mL vial containing 10 mL of deionized water. One set (treatment group) was subjected to heat stress at 45 °C for 1 h in a water bath, while the other set (control group) remained at room temperature. After cooling, 10 mL of deionized water was added to both sets and the vials were incubated at 10 °C for 24 h.

Following incubation, the vials were gently shaken and the initial electrical conductivity (EC) of both the treatment and control groups was measured using a conductivity meter (HI 8033; HANNA Instruments, Portugal). The leaf tissues of both groups were subsequently killed by boiling in a water bath for 30 min and cooled to room temperature, after which the final electrical conductivity was recorded.

Cell membrane injury was calculated as a percentage of injury using the following equation:

where, Ti = initial EC of the heat-stressed sample, Tf = final EC of the heat-stressed sample, Ci = initial EC of the control sample and Cf = final EC of the control sample.

Determination of antioxidant enzyme activities

All antioxidant enzyme analyses were conducted at the Molecular Laboratory of the Plant Breeding Division, Bangladesh Agricultural Research Institute (BARI), Gazipur, Bangladesh. The total soluble protein concentration in the enzyme extracts was determined according to the method of Bradford [8] using bovine serum albumin (BSA) as the standard. The absorbance was measured at 595 nm using a UV–Visible spectrophotometer (Model UV-1800, Shimadzu, Japan). Protein concentrations were calculated from the BSA standard curve and used to express catalase (CAT), ascorbate peroxidase (APX) and glutathione peroxidase (GPX) activities on a protein basis.

Catalase (CAT)

Catalase activity was measured according to Csiszár et al. [10] by monitoring the decrease in absorbance at 240 nm for 1 min caused by H₂O₂ decomposition. The reaction mixture (0.7 mL) contained 50 mM potassium phosphate buffer (pH 7.0), 15 mM H₂O₂ and the enzyme extract. The reaction was initiated by the addition of the enzyme extract and the activity was calculated using an extinction coefficient of 39.4 M^-1 cm^-1.

Ascorbate peroxidase (APX)

Ascorbate peroxidase activity was determined following Nakano and Asada [37]. The 0.7 mL reaction mixture consisted of 50 mM potassium phosphate buffer (pH 7.0), 0.5 mM ascorbate, 0.1 mM H₂O₂, 0.1 mM EDTA and the enzyme extract. The reaction was initiated by the addition of H₂O₂ and the decrease in absorbance at 290 nm for 1 min was recorded. The activity was calculated using an extinction coefficient of 2.8 mM^-1 cm^-1.

Glutathione peroxidase (GPX)

Glutathione peroxidase activity was assayed as described by Elia et al. [13] using H₂O₂ as a substrate. The reaction mixture contained 100 mM sodium phosphate buffer (pH 7.5), 1 mM EDTA, 1 mM NaN₃, 0.12 mM NADPH, 2 mM reduced glutathione (GSH), 1 U of glutathione reductase (GR), 0.6 mM H₂O₂ and 20 µL of enzyme extract. The reaction was initiated by the addition of H₂O₂ and the oxidation of NADPH was monitored at 340 nm for 1 min. The activity was calculated using an extinction coefficient of 6.62 mM^-1 cm^-1.

Yield and yield attributes

Spike number per plant (SPP) were recorded from one representative plant per pot prior to harvest. After harvesting, the number of grains per spike (GPS) was determined from five randomly selected spikes per replication. Thousand-grain weight (TGW) was estimated from 200 hand-counted grains, multiplied by five and adjusted to 12% moisture content (MC).

For yield determination, one plant per pot was harvested at maturity, followed by drying, threshing and cleaning. The grain yield and total aboveground biomass (BM) were recorded on a per-plant basis and the grain yield was expressed at 12% MC with the following formula.

The harvest index (HI) was calculated as the ratio of grain yield to total biomass.

Statistical analysis

The data were analysed using analysis of variance (ANOVA) appropriate for a split-plot randomized complete block design (RCBD) using the Statistical Tool for Agricultural Research [22]. Replication (block) was treated as a random effect, whereas salinity and genotype were treated as fixed effects, with salinity assigned to the main plots and genotype to the subplots. The model included the main effects of salinity and genotype, as well as their interaction (salinity × genotype). The main effect of salinity was tested using the replication × salinity interaction as the main-plot error term, whereas the effects of genotype and the salinity × genotype interaction were tested using the residual subplot error. When ANOVA indicated significant differences, treatment means were separated using Tukey's honestly significant difference (HSD) test at p < 0.05. Confidence interval (CI) and HSD values were computed using STAR. Prior to ANOVA, the assumptions of residual normality and homogeneity of variance were evaluated using the Shapiro–Wilk test and Levene's test, respectively. When the residuals deviated significantly from normality, logarithmic transformation was first evaluated. Where the transformation did not adequately satisfy the normality assumption, sensitivity analyses based on the rank-transformed data were performed using the same split-plot ANOVA model to assess the robustness of the statistical conclusions. The results of the assumption tests and additional analyses are presented in the supplementary materials (Tables S1 and S2). Statistical significance was defined as p < 0.05.

Pearson's correlation analysis was performed separately for each salinity (EC) of the irrigation solution (control, 5 dS m^-1 and 10 dS m^-1) using nine observations (three genotypes × three replications; n = 9) to explore associations among physiological, biochemical and yield-related traits. Correlation matrices were visualized as heatmaps to facilitate pattern recognition. Because the analysis was exploratory, no multiplicity correction was applied and the correlation results were interpreted as descriptive associations rather than as confirmatory evidence of causal or regulatory relationships.

Graphs and data visualizations were produced using multiple platforms: some figures were generated in Microsoft Excel [32] and jamovi [53], while others were created in Python within the Google Colaboratory environment [5] using code developed with assistance from ChatGPT [38].

Integrated physiological efficiency (IPE_ridge) framework

To quantify genotype-specific physiological efficiency under salinity stress, both cost–benefit slope analysis and a ridge regression–based integrated physiological efficiency index (IPE_ridge) were employed.

Standardization of physiological traits

Prior to the cost–benefit slope and ridge regression analyses, all the physiological traits were standardized using Z-scores to remove differences in measurement units and scales and to ensure equal contributions of individual traits to the composite physiological modules. Based on the traits evaluated in this study, four physiological modules were subsequently constructed: water relations (WR; plant water status: LWP, RWC and E), osmotic adjustment (OA; osmolyte accumulation: proline and soluble sugars), oxidative stress (OS; oxidative damage indicators: H₂O₂ and CMI) and antioxidant defence (AD; antioxidant enzyme activities: CAT, APX and GPX). For consistency, the physiological modules were constructed with predefined biological directions. Higher relative water content (RWC), higher transpiration rate (E) and less-negative leaf water potential (LWP) contributed positively to the WR score, whereas higher hydrogen peroxide (H₂O₂) and cell membrane injury (CMI) contributed positively to the OS score. Likewise, higher osmolyte accumulation and antioxidant enzyme activities contributed positively to the OA and AD scores, respectively (Table S4). The raw physiological data are provided in Supplementary excel file 1, from which the standardized physiological module scores (WR, OA, OS and AD) and the corresponding cost–benefit analysis outputs were generated. The calculated outputs are provided in a supplementary file (i.e., ‘Cost–benefit output excel file 1’) and the complete executable python analysis script (i.e., Supplementary file: ‘Cost–benefit analysis python script.py’) is included to facilitate the transparency and reproducibility of the proposed cost–benefit analysis framework. The standardized values were calculated as follows:

where, X_i is the observed value, X̄ is the overall mean of the trait and SD is the standard deviation.

The physiological modules were subsequently derived from the standardized trait values as follows:

where, LWP = leaf water potential, RWC = relative water content and E = transpiration rate.

where, Proline = leaf proline concentration and SS = soluble sugar concentration.

where, H₂O₂ = hydrogen peroxide concentration and CMI = cell membrane injury.

where CAT = catalase activity, APX = ascorbate peroxidase activity and GPX = glutathione peroxidase activity.

Cost–benefit slope analysis (ΔBenefit/ΔCost)

Cost–benefit slope analysis was used to quantify the efficiency with which physiological adjustments were translated into functional outcomes under increasing salinity stress. The standardized physiological modules (WR, OA, OS and AD) were considered cost variables, whereas photosynthetic rate (P_n) and grain yield (GY) were considered benefit variables. The efficiency slope was calculated as follows:

where, B₁ and B₂ are the benefit values at two consecutive salinity levels; C₁ and C₂ are the corresponding cost values and ΔBenefit represents the change in P_n or GY and ΔCost represents the change in WR, OA, OS, or AD. Slopes were calculated separately for the moderate (0 → 5 dS m^-1) and severe (5 → 10 dS m^-1) salinity transitions for each genotype. All cost–benefit slope outputs are provided in a supplementary file (i.e., ‘Cost–benefit output excel file 1’).

This approach is conceptually consistent with established frameworks of trait trade-offs, resource-use efficiency and physiological investment–return relationships in plants [14, 40, 48].

Biological interpretation of slope values: The cost–benefit slope describes the change in functional performance (P_n or GY) relative to the magnitude of physiological adjustment under increasing salinity stress. Positive slope values indicate that physiological changes are associated with gains or maintenance of functional performance, whereas negative values indicate a decline in performance relative to the magnitude of physiological change. Therefore, the biological significance of a slope depends on both its direction and magnitude and was interpreted comparatively among genotypes within each physiological module and salinity transition.

Integrated physiological efficiency (IPE_ridge)

To integrate multiple physiological mechanisms into a single efficiency index, ridge regression was employed to account for multicollinearity among the physiological modules while providing stable coefficient estimates [11, 19]. Four standardized physiological modules, namely, water relations (WR), compatible solute accumulation (OA), oxidative stress (OS) and antioxidant defence (AD) were used as predictor variables. Before model fitting, each physiological module was standardized using z-score transformation so that it had a mean of zero and a standard deviation of one.

Two separate ridge regression models were developed using Python 3.13.5 with the scikit-learn (version 1.8.0) library: one using grain yield (GY) and the other using the photosynthetic rate (Pn) as the response variable. Ridge regression was implemented using the ridge estimator. The ridge penalty parameter (λ; equivalent to alpha in scikit-learn) was optimized using GridSearchCV with leave-one-out cross-validation (LOOCV) over a logarithmically spaced grid of 200 candidate λ values ranging from 10^–4 to 10⁴. The λ value minimizing the cross-validated mean squared error was selected for the final model. The selected λ values, fitted ridge coefficients, intercepts, model performance statistics (R², RMSE and LOOCV RMSE) and the final model equations are presented in Supplementary materials (Table S3). The python analysis script (i.e., Supplementary file: ‘IPE_ridge analysis python script.py’), the standardized physiological module scores used for model fitting (Supplementary excel file 2) and the corresponding ridge regression outputs (Supplementary file: ‘IPE_ridge output excel file 2’) are provided to facilitate the transparency and reproducibility of the proposed IPE_ridge framework.Yield-based IPE_ridge

where GY_i is the grain yield of the ith observation; β₀ is the intercept, β₁–β₄ are the fitted ridge regression coefficients and WR_i, OA_i, OS_i and AD_i are the standardized physiological module scores for the ith observation and ε_i is the residual error.

The integrated physiological efficiency index for grain yield was calculated as follows:

Similarly, a second ridge regression model was developed using the photosynthetic rate as the response variable.

Photosynthesis-based IPE_ridge

where (P_n)_i is the photosynthetic rate of the ith observation.

The corresponding integrated physiological efficiency index was calculated as follows:

Because all the physiological modules were standardized before model fitting, the resulting IPE_ridge values represent relative physiological efficiency with respect to the overall dataset mean and may therefore assume either positive or negative values. Higher IPE_ridge values indicate greater integration of physiological responses with the corresponding functional outcome (photosynthetic rate or grain yield), whereas lower values indicate less effective physiological coordination within the present dataset.

The resulting indices were subsequently analysed using split-plot ANOVA followed by Tukey's HSD test. Their relationships with the corresponding measured response variables (P_n and GY) were further examined using linear regression and model performance was summarized using the coefficient of determination (R²) and root mean square error (RMSE). Because the indices and response variables were derived from the same dataset, these analyses were interpreted as assessments of internal consistency (in-sample fit) rather than independent validation of the proposed index.

Unlike conventional multivariate approaches such as principal component analysis, cluster analysis and correlation-based methods, which primarily summarize patterns of trait variation, the proposed IPE_ridge framework integrates multiple standardized physiological modules into a single quantitative descriptor while accounting for predictor multicollinearity. Accordingly, it was developed as a complementary analytical tool for characterizing physiological coordination in the evaluated genotypes.

Results

Plant water relations (LWP and RWC)

Plant water status, assessed through leaf water potential (LWP) and relative water content (RWC), was significantly affected by salinity (EC: electrical conductivity) of the irrigation solution, genotype and their interaction at both anthesis and 15 days after anthesis (DAA) (p < 0.05; Fig. 1A–B). Both traits decreased progressively with increasing salinity (EC) of the irrigation solution, indicating disruption of the plant water balance under saline irrigation. At anthesis (Fig. 1A), LWP became progressively more negative in a genotype-dependent manner. In BAW 1147, the LWP decreased from − 0.69 MPa under the control to − 0.78 MPa at 10 dS m^-1 irrigation solution (an absolute decrease of 0.09 MPa), whereas the BARI Gom 25 decreased to − 0.85 MPa (absolute decrease of 0.16 MPa) and the BARI Gom 28 to − 0.89 MPa (absolute decrease of 0.21 MPa). A similar trend was observed for RWC, where BAW 1147 maintained relatively high values (87.84% to 80.75%), BARI Gom 25 showed a moderate reduction (88.11% to 77.02%) and BARI Gom 28 declined markedly from 89.21% to 68.10%. These patterns indicate that the ability of BARI Gom 25 to maintain tissue hydration under saline irrigation was intermediate between that of the tolerant and susceptible genotypes.

At 15 DAA (Fig. 1B), the effects of saline irrigation became more pronounced. The LWP became substantially more negative across genotypes, reaching − 1.11 MPa in BAW 1147, − 1.26 MPa in BARI Gom 25 and − 1.46 MPa in BARI Gom 28 at 10 dS m^-1 irrigation solution. Compared with the corresponding control values, these values represented absolute decreases of approximately 0.37, 0.45 and 0.67 MPa, respectively. Similarly, RWC decreased to 71.19%, 62.54% and 50.71% in the respective genotypes, corresponding to reductions of approximately 14–37%. These results indicate that plant water relations become increasingly impaired during grain filling under saline irrigation. Notably, the BARI Gom 25 consistently exhibited intermediate responses between those of BAW 1147 and BARI Gom 28.

Overall, both LWP and RWC decreased progressively with increasing salinity, with BAW 1147 consistently maintaining a less negative LWP and a higher RWC than BARI Gom 25 and BARI Gom 28 did across the salinity treatments.

Compatible solutes

Salinity significantly affected the accumulation of compatible solutes, with significant effects of salinity level, genotype and their interaction observed for both proline and soluble sugar contents (p < 0.05; Fig. 2A–B). Both compounds increased progressively with increasing salinity across all the genotypes, although the magnitude of accumulation differed among the genotypes.

The proline content increased markedly with increasing salinity (EC) of the irrigation solution (Fig. 2A). In BAW 1147, proline increased from 1.00 µmol g^-1 FW under the control to 3.19 µmol g^-1 FW at 5 dS m^-1 and 5.91 µmol g^-1 FW at 10 dS m^-1. Under severe salinity (EC) of the irrigation solution (10 dS m^-1), the corresponding proline concentrations in BARI Gom 25 and BARI Gom 28 were 4.33 and 3.73 µmol g^-1 FW, respectively. Thus, BAW 1147 accumulated the highest proline concentration, whereas BARI Gom 28 accumulated the lowest.

A similar trend was observed for the soluble sugar content (Fig. 2B). In BAW 1147, the amount of soluble sugars increased from 35.93 to 51.49 and 58.31 mg g^-1 DW with increasing salinity. At 10 dS m^-1, the soluble sugar contents reached 56.60 and 51.26 mg g^-1 DW in BARI Gom 25 and BARI Gom 28, respectively. Across the salinity treatments, BAW 1147 consistently maintained the highest soluble sugar content, followed by BARI Gom 25, whereas BARI Gom 28 recorded the lowest values. Overall, both the proline and soluble sugar contents increased progressively with increasing salinity (EC) of the irrigation solution (Fig. 2A–B). Compared with BARI Gom 25 and BARI Gom 28, BAW 1147 consistently accumulated higher concentrations of both compatible solutes across the salinity treatments.

Leaf gas exchange

Salinity (EC) of the irrigation solution significantly affected all the leaf gas exchange traits, with significant effects of salinity (EC) of the irrigation solution, genotype and their interaction observed for the net photosynthetic rate (P_n), stomatal conductance (g_sc), intercellular CO₂ concentration (C_i), transpiration rate (E) and water use efficiency (WUE) (p < 0.05; Fig. 3A–E). Increasing salinity (EC) of the irrigation solution from the control to 10 dS m^-1 progressively reduced P_n, g_sc, C_i and E across all the genotypes, although the magnitude of these responses differed among the genotypes. In contrast, WUE increased with increasing the salinity (EC) of the irrigation solution. These results indicate genotype-dependent differences in leaf gas exchange responses to salinity (EC) of the irrigation solution.

The net photosynthetic rate (P_n) decreased progressively with increasing salinity (EC) of the irrigation solution (Fig. 3A). Although the genotypes differed significantly under control conditions, the relative reduction at 10 dS m^-1 irrigation solution was smallest in BAW 1147 (15.5%), followed by BARI Gom 25 (19.0%) and BARI Gom 28 (25.5%). Accordingly, BAW 1147 maintained the highest P_n (16.09 µmol CO₂ m^-2 s^-1), whereas BARI Gom 25 and BARI Gom 28 recorded lower values of 13.79 and 13.59 µmol CO₂ m^-2 s^-1, respectively.

Stomatal conductance (g_sc) also decreased with increasing salinity (EC) of the irrigation solution (Fig. 3B). At 10 dS m^-1 irrigation solution, BAW 1147 maintained a higher g_sc (0.129 mmol m^-2 s^-1) than BARI Gom 25 (0.113 mmol m^-2 s^-1) and BARI Gom 28 (0.101 mmol m^-2 s^-1). Relative to the control, g_sc declined by approximately 29.2%, 35.3% and 50.0% in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively.

The intercellular CO₂ concentration (C_i) also decreased with increasing salinity (EC) of the irrigation solution (Fig. 3C). At 10 dS m^-1 irrigation solution, the C_i values were 216.22, 204.30 and 186.18 µmol CO₂ mol^-1 in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively. In contrast to the differences in the other gas exchange traits, genotypic differences under control conditions were not significant. However, the relative reduction was smallest in BAW 1147 (10.9%) and greatest in BARI Gom 28 (31.6%).

The transpiration rate (E) decreased progressively with increasing salinity (EC) of the irrigation solution (Fig. 3D). At 10 dS m^-1 irrigation solution, BARI Gom 28 maintained a slightly higher E (3.98 mmol H₂O m^-2s^-1) than BARI Gom 25 (3.60 mmol H₂O m^-2 s^-1) and BAW 1147 (3.56 mmol H₂O m^-2 s^-1). Relative reductions in E were comparable among genotypes, ranging from 29.2% to 32.6%.

In contrast, water use efficiency (WUE = P_n/E) increased with increasing salinity (EC) of the irrigation solution (Fig. 3E). At 10 dS m^-1 irrigation solution, BAW 1147 had the highest WUE (4.52 µmol CO₂ mmol^-1 H₂O), followed by BARI Gom 25 (3.84 µmol CO₂ mmol^-1 H₂O) and BARI Gom 28 (3.41 µmol CO₂ mmol^-1 H₂O). Relative to the control, WUE increased by approximately 25.4%, 15.4% and 5.2% in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively.

Overall, salinity (EC) of the irrigation progressively affected all the measured leaf gas exchange traits (Fig. 3A–E). Across the salinity treatments, compared with the other two genotypes, BAW 1147 consistently maintained higher P_n, g_sc, C_i and WUE, whereas BARI Gom 28 resulted in the greatest reduction in these traits.

Photosynthetic pigments

Salinity (EC) of the irrigation solution significantly affected all the measured photosynthetic pigments, with significant effects of salinity level, genotype and their interaction observed for chlorophyll a (Chl a), chlorophyll b (Chl b), the Chl a/Chl b ratio and total chlorophyll content (p < 0.05; Fig. 4A–E). Increased salinity (EC) of the irrigation solution progressively reduced the Chl a, Chl b, total chlorophyll and carotenoids contents across all the genotypes, whereas the Chl a/Chl b ratio remained relatively stable. Chl a decreased progressively with increasing salinity (EC) of the irrigation solution (Fig. 4A). At 10 dS m^-1 irrigation solution, the Chl a concentration decreased from 1.71 to 1.54 mg g^-1 FW in BAW 1147, whereas the values of the BARI Gom 25 and BARI Gom 28 were lower (1.41 and 1.27 mg g^-1 FW, respectively). Compared with that in the control, the Chl a concentration in the BAW 1147, BARI Gom 25 and BARI Gom 28 decreased by approximately 10.1%, 13.8% and 23.8%, respectively.

A similar trend was observed for Chl b (Fig. 4B). At 10 dS m^-1 irrigation solution, the Chl b concentrations were 1.04, 0.98 and 0.96 mg g^-1 FW in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively. The relative reduction was smallest in BAW 1147 and greatest in BARI Gom 28.

In contrast, the Chl a/Chl b ratio remained relatively stable across the salinity treatments (Fig. 4C), with only a slight decline observed in the BARI Gom 28, indicating proportional decreases in both chlorophyll fractions. The total chlorophyll content followed the combined pattern of Chl a and Chl b (Fig. 4D). At 10 dS m^-1 irrigation solution, the total chlorophyll concentration decreased from 2.86 to 2.58 mg g^-1 FW in BAW 1147, whereas the BARI Gom 25 and BARI Gom 28 declined further to 2.39 and 2.23 mg g^-1 FW, respectively, corresponding to relative reductions of approximately 9.8%, 13.3% and 21.1%, respectively.

The carotenoids content also decreased progressively with increasing salinity (EC) of the irrigation solution (Fig. 4E). Relative to the respective control, carotenoids content decreased by approximately 35.3%, 49.6% and 32.5% in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively. Thus, the greatest relative reduction in the carotenoids content occurred in the BARI Gom 25. Under severe salinity (EC) of the irrigation solution (10 dS m^-1), however, BAW 1147 maintained the highest absolute carotenoids concentration (2.13 mg g^-1 FW), followed by BARI Gom 28 (1.90 mg g^-1 FW) and BARI Gom 25 (1.64 mg g^-1 FW). These findings show that the ranking of genotypes based on percentage reduction differed from that based on the absolute concentration of carotenoids maintained under salinity.

Overall, salinity (EC) of the irrigation solution progressively reduced the leaf pigment contents across all the genotypes (Fig. 4A–E). BAW 1147 consistently maintained higher chlorophyll and carotenoids concentrations under saline conditions than BARI Gom 25 and BARI Gom 28. BARI Gom 28 generally exhibited the lowest chlorophyll concentrations under salinity, whereas BARI Gom 25 showed the greatest relative reduction in carotenoids content.

Oxidative stress responses (H₂O₂ and cell membrane injury)

Salinity (EC) of the irrigation solution significantly affected oxidative stress-related traits, with significant effects of salinity (EC) of the irrigation solution, genotype and their interaction observed for hydrogen peroxide (H₂O₂) concentration and cell membrane injury (CMI) (p < 0.05; Fig. 5A–B). Both the H₂O₂ concentration and the CMI increased progressively with increasing salinity (EC) of the irrigation solution across all the genotypes, although the magnitude of these responses differed among the genotypes. The H₂O₂ concentration increased consistently with increasing salinity (EC) of the irrigation solution (Fig. 5A). The mean H₂O₂ concentration increased from approximately 12.6 µmol g^-1 FW under the control to 15.2 µmol g^-1 FW at 5 dS m^-1 irrigation solution and further to approximately 17.0 µmol g^-1 FW at 10 dS m^-1 irrigation solution. At the highest salinity level (EC) of the irrigation solution, BAW 1147 maintained the lowest H₂O₂ concentration (approximately 13.4 µmol g^-1 FW), followed by BARI Gom 25 (approximately 15.8 µmol g^-1 FW), whereas BARI Gom 28 recorded the highest concentration (approximately 17.0 µmol g^-1 FW).

A similar trend was observed for cell membrane injury (CMI) (Fig. 5B). The mean CMI increased from approximately 41% under control conditions to 52% at 5 dS m^-1 and 61% at 10 dS m^-1. At 10 dS m^-1 irrigation solution, BAW 1147 had the lowest CMI (approximately 50%), followed by BARI Gom 25 (approximately 55%), whereas BARI Gom 28 had the highest degree of membrane injury (approximately 61%). The CMI pattern closely corresponded with the observed changes in H₂O₂ concentration across genotypes.

Overall, both the H₂O₂ concentration and CMI increased progressively with increasing salinity (EC) of the irrigation solution in all the genotypes (Fig. 5A–B). Across the salinity treatments, BAW 1147 consistently maintained lower H₂O₂ concentration and CMI than BARI Gom 25 and BARI Gom 28, whereas BARI Gom 28 exhibited the highest values for both traits.

Antioxidant enzyme activities (CAT, APX and GPX)

Salinity (EC) of the irrigation solution significantly affected antioxidant enzyme activity, with significant effects of salinity (EC) of the irrigation solution, genotype and their interaction observed for catalase (CAT), ascorbate peroxidase (APX) and glutathione peroxidase (GPX) (p < 0.05; Fig. 6A–C). Additional analyses performed for CAT and GPX to assess the effect of departures from residual normality produced conclusions consistent with those obtained from the original ANOVA (Table S2). The activities of all three enzymes increased progressively with increasing salinity (EC) of the irrigation solution, although the magnitude of the response differed among the genotypes.

CAT activity increased with increasing salinity (EC) of the irrigation solution (Fig. 6A), increasing from approximately 110 µmol min^-1 mg^-1 protein under control conditions to approximately 117 µmol min^-1 mg^-1 protein at 5 dS m^-1 irrigation solution and 125 µmol min^-1 mg^-1 protein at 10 dS m^-1 irrigation solution. Compared with that of the control, the CAT activity increased by 9.95%, 8.51% and 2.92% at 10 dS m^-1 in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively. Under severe salinity (EC) of the irrigation solution, BAW 1147 had the highest CAT activity (approximately 125 µmol min^-1 mg^-1 protein), followed by BARI Gom 25 (approximately 116 µmol min^-1 mg^-1 protein) and BARI Gom 28 (approximately 114 µmol min^-1 mg^-1 protein). A similar trend was observed for APX activity (Fig. 6B), which increased from approximately 0.49–0.56 µmol min^-1 mg^-1 protein under control conditions to approximately 0.59–0.62 µmol min^-1 mg^-1 protein at 10 dS m^-1 irrigation solution. Compared with the control, APX activity increased by 18.40%, 10.92% and 6.52% in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively. Although BARI Gom 25 recorded the highest absolute APX activity at 10 dS m^-1 irrigation solution (approximately 0.62 µmol min^-1 mg^-1 protein), the greatest relative increase was observed in BAW 1147.

GPX activity showed the greatest increase with increasing salinity (EC) of the irrigation solution (Fig. 6C). The activity increased from approximately 91–94 µmol min^-1 mg^-1 protein under control conditions to approximately 120 µmol min^-1 mg^-1 protein at 10 dS m^-1 irrigation solution. At this salinity level, BAW 1147 exhibited the highest GPX activity (approximately 120 µmol min^-1 mg^-1 protein), followed by BARI Gom 25 (approximately 103 µmol min^-1 mg^-1 protein) and BARI Gom 28 (approximately 100 µmol min^-1 mg^-1 protein).

Taken together, the CAT, APX and GPX activities increased progressively with increasing salinity (EC) of the irrigation solution in all the genotypes (Fig. 6A–C). Across salinity treatments, BAW 1147 generally maintained higher antioxidant enzyme activities than BARI Gom 25 and BARI Gom 28, whereas BARI Gom 28 exhibited the lowest enzyme activities under severe salinity (EC) of the irrigation solution.

Grain yield and yield components

Salinity significantly affected grain yield and all measured yield-related traits, with significant effects of salinity level, genotype and their interaction (p < 0.05; Fig. 7; Table 1). The grain yield decreased progressively with increasing salinity in all the genotypes, although the magnitude of the reduction differed among the genotypes.

Fig. 7 — Interactive effects of salinity (EC) of the irrigation solution (control, 5 dS m^-1 and 10 dS m^-1) and wheat genotype (BAW 1147, BARI Gom 25 and BARI Gom 28) on grain yield (GY). The bars represent the means with 95% confidence intervals (CIs) of three replicates (n = 3). Different letters indicate significant differences among salinity (EC) of the irrigation solution × genotype interaction means according to Tukey’s honestly significant difference (HSD) test at p < 0.05

Table 1.

Interactive effects of salinity (EC) of the irrigation solution (control, 5 dS m^-1 and 10 dS m^-1) and wheat genotype (BAW 1147, BARI Gom 25 and BARI Gom 28) on yield traits

Salinity	Genotype	Spike plant⁻¹ (no.)	Grain spike⁻¹ (no.)	TGW (g)	Biomass (g plant⁻¹)	Harvest index
Control	BAW 1147	9.17 a	57.63 a	58.76 a	46.25 a	0.43 a
	BARI Gom 25	8.33 b	54.73 b	53.39 c	43.21 c	0.43 a
	BARI Gom 28	9.50 a	54.30 b	49.71 e	46.50 a	0.42 a
5 dS m⁻¹	BAW 1147	8.67 a	55.67 b	57.34 b	44.92 b	0.42 a
	BARI Gom 25	8.00 b	52.50 c	52.03 d	40.30 d	0.40 b
	BARI Gom 28	8.33 b	51.53 d	47.05 f	41.15 d	0.38 b
10 dS m⁻¹	BAW 1147	8.30 b	53.73 c	53.07 c	42.72 c	0.40 b
	BARI Gom 25	7.50 b	49.70 e	47.38 f	38.56 e	0.39 b
	BARI Gom 28	7.30 b	47.50 f	41.84 g	37.90 e	0.36 c

Open in a new tab

The values are the means (n = 3). Means within a column followed by the same letter are not significantly different according to Tukey’s HSD test at p < 0.05

TGW thousand-grain weight

At 10 dS m^-1 irrigation solution (Fig. 7), BAW 1147 resulted in the highest grain yield (17.27 g plant^-1), which was approximately 88% of its control yield (19.70 g plant^-1). The BARI Gom 25 showed an intermediate reduction to 15.11 g plant^-1 (approximately 82% of the control), whereas the BARI Gom 28 exhibited the greatest reduction to 13.57 g plant^-1 (approximately 70% of the control). At 5 dS m^-1 irrigation solution, yield reductions were relatively small, with BAW 1147 and BARI Gom 25 retaining approximately 90% of their respective control yields, whereas BARI Gom 28 showed a greater reduction.

The yield components exhibited similar responses to increasing salinity (Table 1). Under 10 dS m^-1 irrigation solution, the spike number per plant decreased by 9.5%, 10.0% and 23.2% in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively. Despite these reductions, BAW 1147 retained the highest spike number per plant (8.30), followed by BARI Gom 25 (7.50) and BARI Gom 28 (7.30). The number of grains per spike also decreased with increasing salinity (EC) of the irrigation solution, with reductions of 6.8%, 9.2% and 12.5% in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively. At 10 dS m^-1 irrigation solution, BAW 1147 retained more than 53 grains per spike, whereas BARI Gom 28 recorded approximately 47.5 grains per spike. Thousand-grain weight (TGW) also decreased under salinity (EC) of the irrigation solution, decreasing by 9.7%, 11.3% and 15.8% in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively. Nevertheless, BAW 1147 recorded the highest TGW (approximately 53.1 g), whereas BARI Gom 28 recorded the lowest value (41.84 g). The biomass per plant also decreased, with relative reductions of 7.6%, 10.8% and 18.5% in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively. The harvest index was relatively less affected by salinity (EC) of the irrigation solution, with BAW 1147 having the highest value (0.40), followed by BARI Gom 25 (0.39), whereas BARI Gom 28 had the lowest value (0.36).

Overall, the grain yield and all measured yield components decreased progressively with increasing salinity (Fig. 7; Table 1). Across the salinity treatments, BAW 1147 consistently recorded higher grain yield and yield-component values than BARI Gom 25 and BARI Gom 28, whereas BARI Gom 28 exhibited the largest reductions.

Correlation dynamics of grain yield under salinity stress

Correlation analysis (Figs. 8 & 9) suggested progressive changes in trait associations with increasing salinity (EC) of the irrigation solution, reflected by increasing colour intensity in the correlation heatmaps. Under control conditions, grain yield (GY) was primarily associated with biomass production and photosynthetic traits, whereas its associations with biochemical traits were relatively weak (Figs. 8A & 9A).

Fig. 8 — Correlation heatmaps between grain yield (GY) and related physiological and biochemical traits of wheat under (A) control, (B) 5 dS m^-1 irrigation solution and (C) 10 dS m^-1 irrigation solution. Pearson's correlation coefficients (r) were calculated using nine observations (n = 9) for each treatment. The values inside the cells represent Pearson’s correlation coefficients (r), with significance levels denoted as * (p ≤ 0.05), ** (p ≤ 0.01) and ns (nonsignificant). The colour intensity indicates the correlation strength, with deep green denoting a strong positive correlation

Fig. 9 — Correlation heatmaps of wheat yield and yield-related traits under three salinity (EC) of the irrigation solution: A Control, B 5 dS m^-1 irrigation solution and C 10 dS m^-1 irrigation solution. Pearson's correlation coefficients (r) were calculated using nine observations (n = 9) for each treatment. The values inside the cells represent Pearson’s correlation coefficients (r), with significance levels denoted as * (p ≤ 0.05), ** (p ≤ 0.01) and ns (nonsignificant). Deeper green shades indicate stronger positive correlations, whereas red shades represent negative correlations; pale colours correspond to weak or nonsignificant relationships. Because each correlation analysis was based on a limited sample size (n = 9), the observed correlations should be interpreted as exploratory associations rather than validated physiological relationships

At moderate salinity (EC) of the irrigation solution (5 dS m^-1), the correlation patterns became more apparent, with grain yield showing positive associations with biomass, the harvest index, grain number per spike, thousand-grain weight, water-use efficiency, chlorophyll content, compatible solutes and antioxidant enzyme activities (Figs. 8B & 9B). Negative associations between grain yield and oxidative stress indicators (H₂O₂ and CMI) were more apparent under the 5 dS m^-1 irrigation solution treatment.

Under severe salinity (EC) of the irrigation solution (10 dS m^-1), stronger positive correlations were observed between grain yield and grain number per spike, thousand-grain weight, harvest index, water-use efficiency, chlorophyll content, compatible solutes and antioxidant enzyme activities, whereas stronger negative associations were observed with H₂O₂ and CMI (Figs. 8C & 9C). Overall, the heatmaps suggested increasingly coordinated association patterns among physiological, biochemical and yield-related traits with increasing salinity (EC) of the irrigation solution.

Overall, the correlation patterns suggest a transition from predominantly biomass-associated yield relationships under nonstress conditions to broader multitrait associations under salinity. Because Pearson's correlation analysis was performed using nine observations within each salinity level (n = 9) and no multiplicity correction was applied, these findings should be considered exploratory associations rather than evidence of causal or validated physiological relationships. Nevertheless, they illustrate the potential value of integrated trait-based approaches for characterizing salinity responses in the evaluated wheat genotypes.

Cost–benefit relationships and integrated physiological efficiency

Cost–benefit slope analysis for P_n and GY

The cost–benefit slopes varied among the physiological modules, genotypes and salinity transitions (Fig. 10). For both the photosynthetic rate (P_n) and grain yield (GY), the WR-related slopes were consistently positive, whereas the OA-, OS- and AD-related slopes were generally negative. During the 0 → 5 dS m^-1 transition, BAW 1147 generally exhibited less negative OA-, OS- and AD-related slopes for both P_n and GY than the other genotypes, whereas BARI Gom 25 showed intermediate values and BARI Gom 28 displayed the most negative slopes in most cases. In contrast, the WR-related slopes were positive in all the genotypes, indicating positive associations between water-relations adjustment and both photosynthetic and yield performance during the initial phase of salinity stress.

As the salinity increased from 5 to 10 dS m^-1 irrigation solution, the genotype-specific patterns became more differentiated. BAW 1147 maintained the highest WR-related slopes for both P_n and GY and these slopes increased relative to the 0 → 5 dS m^-1 transition. In contrast, the WR-related slopes decreased substantially in BARI Gom 25 and decreased for the GY in the BARI Gom 28. Although the BARI Gom 25 exhibited less negative OA-, OS- and AD-related slopes for P_n and less negative OA- and OS-related slopes for GY than BAW 1147, these responses were accompanied by reduced WR-related benefits. In comparison, BAW 1147 maintained the highest WR-related slopes for both P_n and GY together with the least negative AD → GY slope, indicating a more favourable balance between physiological benefits and costs under severe salinity. The BARI Gom 28 continued to exhibit strongly negative OA-, OS- and AD-related slopes for both P_n and GY, although OS → P_n and OS → GY became less negative than BAW 1147 during the 0 → 5 dS m^-1 transition. Overall, increasing salinity altered the balance between water-relations benefits and physiological costs among genotypes, with BAW 1147 maintaining greater WR-related benefits and lower AD-related yield costs under severe salinity, BARI Gom 25 exhibiting comparatively lower OA-, OS- and AD-related costs and BARI Gom 28 generally showing the least favourable cost–benefit relationships.

Integrated physiological efficiency (IPE_ridge) and its association with grain yield and photosynthetic performance

Salinity (EC) of the irrigation solution significantly affected IPE_ridge(GY), with significant effects of salinity (EC) of the irrigation solution, genotype and their interaction (p < 0.05; Fig. 11A). Under control conditions, all three genotypes exhibited positive IPE_ridge(GY) values, with BARI Gom 28 (2.32), BAW 1147 (2.12) and BARI Gom 25 (1.69), although BARI Gom 28 and BAW 1147 did not differ significantly. With increasing salinity (EC) of the irrigation solution, IPE_idge(GY) declined progressively in all genotypes, although the magnitude of the decline differed among them. At 5 dS m^-1 irrigation solution, BAW 1147 maintained the highest IPE_ridge(GY) value (0.90), whereas BARI Gom 25 and BARI Gom 28 declined to − 0.54 and − 0.98, respectively. Under 10 dS m^-1 irrigation solution, BAW 1147 maintained the highest value (0.17), whereas BARI Gom 25 and BARI Gom 28 declined further to − 1.84 and − 3.84, respectively. These results indicate that BAW 1147 maintained greater integrated physiological efficiency than the other two genotypes under increasing salinity.

A strong positive in-sample association was observed between IPE_ridge(GY) and grain yield (GY = 17.15 + 1.05 × IPE_ridge(GY), R² = 0.911, p < 0.001; Fig. 11B). Because IPE_ridge(GY) was derived using grain yield as the response variable, this relationship represents the expected internal consistency (in-sample fit) of the proposed index within the present dataset rather than independent validation.

A similar pattern was observed for the photosynthesis-based index (Fig. S1A). IPE_ridge(Pn) was significantly affected by salinity (EC) of irrigation solution, genotype and their interaction (p < 0.05). Under control conditions, all three genotypes exhibited positive IPE_ridge(P_n) values, with BARI Gom 28 (2.19), BAW 1147 (1.87) and BARI Gom 25 (1.29). With increasing salinity, IPE_ridge(P_n) decreased progressively in all genotypes. At 5 dS m^-1 irrigation solution, BAW 1147 maintained the highest value (0.95), whereas BARI Gom 28 and BARI Gom 25 declined to − 0.37 and − 0.42, respectively. Under 10 dS m^-1 irrigation solution, IPE_ridge(P_n) decreased further to − 0.35, − 1.94 and − 3.22 in BAW 1147, BARI Gom 25 and BARI Gom 28, respectively (Fig. S1A). A strong positive in-sample association was also observed between IPE_ridge(P_n) and the photosynthetic rate (P_n = 13.48 + 1.08 × IPE_ridge(P_n), R² = 0.847, p < 0.001; Fig. S1B). As with the grain yield-based index, this relationship reflects the expected internal consistency of the ridge regression model within the analyzed dataset rather than independent validation.

Collectively, the IPE_ridge values decreased progressively with increasing salinity (EC) of the irrigation solution in both models, with BAW 1147 consistently maintaining higher values than BARI Gom 25 and BARI Gom 28 across salinity (EC) of the irrigation solution (Fig. 11 and Fig. S1).

Discussion

Salinity is a complex environmental constraint that simultaneously imposes osmotic, ionic and oxidative stresses, collectively disrupting plant water relations, photosynthetic function, metabolism and reproductive development [35, 55]. Because these processes interact rather than operate independently, salinity tolerance is generally considered to result from coordinated physiological adjustment rather than the response of any single trait. Within this framework, the present study compared three contrasting wheat genotypes to determine how multiple physiological and biochemical responses were associated with the maintenance of grain yield under increasing salinity (EC) of the irrigation solution. The results consistently showed that BAW 1147 exhibited comparatively smaller reductions in water status, photosynthetic performance, pigment content and yield, together with greater accumulation of compatible solutes and antioxidant enzyme activities than BARI Gom 25 and BARI Gom 28. These observations suggest that salinity tolerance in the evaluated genotypes was associated with coordinated regulation of several physiological processes rather than any individual response. Nevertheless, because only three genotypes were evaluated under controlled conditions, these findings should be interpreted within the scope of the present dataset rather than as universal determinants of wheat salinity tolerance.

The maintenance of plant water status appeared to be one of the earliest responses associated with salinity tolerance. Reduced external osmotic potential restricts water uptake by roots, lowers leaf turgor and consequently limits cell expansion, stomatal opening and carbon assimilation [24, 35, 54]. Consistent with this physiological framework, BAW 1147 maintained a less negative leaf water potential and higher relative water content than the other two genotypes, whereas BARI Gom 28 exhibited the greatest decline in both parameters (Fig. 1). These differences were accompanied by relatively small reductions in gas exchange and grain yield in BAW 1147, suggesting that improved maintenance of tissue hydration was associated with sustained physiological activity under saline irrigation conditions. Similar relationships between the maintenance of leaf water status and improved salinity tolerance have been reported in wheat and other cereal crops [18, 21]. However, the present results indicate association rather than causation because the individual contribution of water relations to yield maintenance was not experimentally separated from other concurrent physiological responses.

The greater accumulation of proline and soluble sugars observed in BAW 1147 was also consistent with improved physiological adjustment to salinity (EC) of the irrigation solution (Fig. 2). Compatible solutes are widely recognized for their contribution to osmotic balance, stabilization of proteins and membranes, protection of cellular structures and maintenance of enzyme function under abiotic stress [3, 26, 50, 51]. Their greater accumulation in the tolerant genotype therefore suggests that osmotic regulation formed part of a broader adaptive response to increasing salinity. Nevertheless, compatible solute accumulation alone should not be regarded as direct evidence of osmotic adjustment because the osmotic potential and the quantitative contribution of individual solutes were not measured in the present study. Moreover, the relatively small differences among genotypes indicate that proline and soluble sugars probably acted together with other physiological processes rather than functioning as independent determinants of tolerance. Similar integrated relationships among water status, compatible solutes, pigment stability, oxidative responses and antioxidant metabolism have recently been described in cotton exposed to drought stress [29]. Although that study involved a different crop and stress condition, it provides a useful cross-crop methodological comparison supporting the simultaneous evaluation of water status, pigments, osmolytes and antioxidant responses rather than direct evidence for wheat salinity tolerance. The maintenance of leaf gas exchange provides an important physiological link between water status and grain production. As salinity increased (EC) of the irrigation solution, the reduction in stomatal conductance was accompanied by a lower intercellular CO₂ concentration and a reduced net photosynthetic rate, indicating the progressive restriction of carbon assimilation. Such responses are characteristic of salinity-induced stomatal limitation, where partial stomatal closure conserves water but simultaneously restricts CO₂ diffusion into the leaf [14]. Compared with the susceptible genotype, the BAW 1147 exhibited comparatively smaller reductions in stomatal conductance and photosynthetic rate together with a greater increase in water use efficiency (Fig. 3). These observations suggest that the tolerant genotype maintained a more favourable balance between carbon gain and water loss under saline irrigation conditions. In contrast, the BARI Gom 28 showed greater reductions in photosynthetic activity and only limited improvement in water use efficiency, indicating that photosynthetic performance became increasingly constrained as stress intensified. Collectively, these responses are consistent with the view that the maintenance of gas exchange contributed to sustained biomass accumulation and reproductive development under saline irrigation condition, although the present experiment was not designed to partition stomatal and nonstomatal limitations to photosynthesis independently.

Photosynthetic pigment retention further supported the maintenance of photosynthetic function. Salinity-induced reductions in chlorophyll content may result from the inhibition of pigment biosynthesis, enhanced degradation, disruption of chloroplast ultrastructure and oxidative injury to the photosynthetic apparatus. BAW 1147 consistently retained higher chlorophyll concentrations than the other genotypes, whereas BARI Gom 28 exhibited the greatest pigment loss (Fig. 4). Carotenoids are particularly important because they participate not only in light harvesting but also in thermal energy dissipation and the scavenging of reactive oxygen species generated during stress [1]. An important observation of the present study was that genotype ranking differed depending on whether carotenoids were expressed as absolute concentrations under stress or as percentage reductions relative to genotype-specific controls. Although the BARI Gom 25 exhibited the greatest relative decline, BAW 1147 maintained the highest absolute carotenoids concentration under severe salinity (EC) of the irrigation solution (Fig. 4). This distinction highlights the importance of considering both baseline pigment status and stress-induced changes when comparing stress responses among genotypes. Overall, the greater retention of chlorophylls and carotenoids in BAW 1147 was associated with higher photosynthetic performance under saline irrigation conditions, although the present data do not demonstrate that pigment stability alone accounted for the observed differences in grain yield.

The coordinated responses of oxidative stress indicators and antioxidant enzymes further highlight the contrasting physiological behaviour of the evaluated genotypes under saline irrigation conditions. Salinity enhances the generation of reactive oxygen species (ROS), including H₂O₂, which, when produced beyond the antioxidant buffering capacity of plant cells, can disrupt membrane integrity, impair chloroplast function, accelerate pigment degradation and ultimately reduce photosynthetic efficiency [33, 43, 47]. In the present study, progressive increases in H₂O₂ concentration and cell membrane injury were accompanied by increased CAT, APX and GPX activity. However, the magnitude of oxidative damage differed among genotypes. BAW 1147 consistently exhibited comparatively lower H₂O₂ accumulation (Fig. 5) and membrane injury together with greater antioxidant enzyme activity (Fig. 6) than BARI Gom 25 and BARI Gom 28. These observations are consistent with a more favourable oxidative balance under salinity (EC) of the irrigation solution. Nevertheless, antioxidant enzyme activity alone should not be interpreted as evidence of greater tolerance because elevated enzyme activity may also reflect increased oxidative stress. Instead, the combined occurrence of relatively high antioxidant activities, lower H₂O₂ accumulation and reduced membrane injury provides stronger evidence of coordinated oxidative regulation than any individual biochemical parameter.

Increasing evidence indicates that ROS function not only as damaging molecules but also as important signalling components regulating plant stress adaptation. Consequently, stress tolerance depends on maintaining ROS within an appropriate physiological range rather than eliminating them completely [33, 43]. A similar concept was proposed by Khan et al. [30], who applied an oxidative-window framework to show that drought- and heat-resilient cotton genotypes maintained physiological performance through coordinated ROS regulation and antioxidant homeostasis rather than maximal antioxidant enzyme activity alone. Although that study involved a different crop and stress combination, it provides cross-crop conceptual support for interpreting antioxidant activity together with ROS accumulation and oxidative damage, rather than as direct evidence for wheat. Similarly, the results of the present study suggest that the lower oxidative damage observed in BAW 1147 was associated with more effective coordination between ROS accumulation and antioxidant activity under saline irrigation conditions.

The grain yield represents the cumulative outcome of multiple physiological and reproductive processes rather than the direct consequence of any individual physiological trait. While the spike number per plant, grain number per spike, thousand-grain weight, biomass production and harvest index directly determine grain yield, physiological traits such as water relations, photosynthetic performance, pigment stability, compatible solute accumulation and antioxidant activity contribute indirectly by supporting assimilate production and reproductive development. Under saline irrigation conditions, reductions in tissue hydration and photosynthetic activity are expected to decrease carbon assimilation, whereas oxidative damage may further impair chloroplast function and membrane stability. Consequently, the maintenance of these physiological processes is likely to support spike survival, grain set and grain filling, thereby contributing indirectly to grain yield. The relatively small reductions in spike number per plant, grain number per spike, thousand-grain weight, biomass and harvest index observed in BAW 1147 (Table 1) were therefore consistent with its comparatively favourable physiological performance. However, because all the traits were measured simultaneously rather than manipulated experimentally, the present study identifies coordinated associations rather than direct causal relationships between physiological responses and yield formation. Accordingly, the maintenance of grain yield under salinity is more appropriately interpreted as the integrated outcome of multiple complementary physiological processes than as the consequence of any single trait.

Exploratory correlation analysis further supports this interpretation. Under control condition, grain yield was primarily associated with biomass production (Fig. 9) and photosynthetic traits (Fig. 8), whereas under increasing salinity, the association patterns progressively expanded to include water use efficiency, compatible solutes, pigment content, antioxidant enzyme activity and oxidative stress indicators (Fig. 8). This shift suggests that the maintenance of grain yield under salinity became associated with increasingly interconnected physiological and biochemical responses rather than a single dominant trait. Recent cotton studies have similarly integrated morpho-physiological, biochemical, oxidative-balance and multivariate traits to improve genotype discrimination under salinity and drought [16, 17, 29]. Although these studies involved different crops and stress conditions, they provide methodological support for integrated trait-based characterization rather than direct evidence for physiological relationships in wheat. Nevertheless, because the correlation analyses in the present study were based on only nine observations within each salinity level and no adjustment for multiple comparisons was applied, these relationships should be regarded as exploratory descriptions of association patterns rather than evidence of causal interactions or regulatory networks.

The cost–benefit slope analysis and the IPE_ridge framework were developed to complement conventional physiological interpretation by integrating multiple correlated physiological responses into quantitative descriptors of genotype performance under salinity. Rather than evaluating individual traits independently, these analytical approaches summarize how coordinated changes in water relations, compatible solute accumulation, oxidative stress and antioxidant defence were associated with photosynthetic performance and grain yield across increasing salinity. Because salinity tolerance involves the simultaneous interaction of multiple physiological processes, such integrative analyses may provide additional insight beyond the interpretation of individual traits alone.

The cost–benefit slope analysis revealed that the relative contributions of different physiological modules changed with increasing salinity and differed among genotypes. Positive WR-related slopes indicated that changes in the standardized water-relations module were positively associated with changes in the photosynthetic rate and grain yield, whereas negative OA-, OS- and AD-related slopes reflected inverse relationships defined by the fitted analytical framework. These analytical descriptors should not be interpreted as direct metabolic costs, energetic expenditures, or resource-allocation trade-offs because the terms “cost” and “benefit” represent operational relationships within the proposed framework rather than experimentally measured physiological quantities. Nevertheless, the genotype-specific differences observed across salinity transitions suggest that the relative associations among water relations, osmotic adjustment, oxidative stress, antioxidant responses and functional performance changed as salinity intensified. In particular, BAW 1147 generally maintained higher WR-related slopes together with comparatively less negative AD → GY slopes under severe salinity (EC) of the irrigation solution, whereas BARI Gom 28 exhibited less favourable slope patterns across most physiological modules (Fig. 10). These observations are consistent with stronger physiological integration in the tolerant genotype but should not be interpreted as evidence of causal interactions among the evaluated modules.

The IPE_ridge framework extends this integrative approach by combining correlated physiological modules into response-weighted indices derived through ridge regression. Unlike principal component analysis (PCA), cluster analysis, or correlation-based approaches, which primarily summarize variation among traits or classify genotypes according to similarity, ridge regression estimates the relative contribution of correlated predictor variables to a predefined biological response while reducing coefficient instability arising from multicollinearity [11, 19]. Consequently, PCA and cluster analysis describe patterns of variation within a dataset, whereas IPE_ridge provides a quantitative descriptor of the combined association between standardized physiological modules and a defined functional response, namely, the photosynthetic rate or grain yield. The comparatively higher IPE_ridge values observed in BAW 1147 under salinity (EC) of the irrigation solution were therefore consistent with stronger integrated physiological performance than that observed in the other evaluated genotypes (Fig. 11).

Importantly, the strong relationships between IPE_ridge and its corresponding response variables should be interpreted appropriately. Because the index was derived using the photosynthetic rate or grain yield and subsequently regressed against the same response variable, the resulting relationships represent the expected in-sample associations and demonstrate the internal consistency of the analytical framework within the present dataset rather than independent predictive validation. Consequently, the regression analyses should not be interpreted as evidence that IPE_ridge has predictive capability for independent populations or environments. Independent validation across additional genotype panels, growing environments and seasons is needed before broader application of the framework can be established.

The principal novelty of the present study therefore lies not in measuring individual physiological or biochemical traits, which have been widely investigated in wheat salinity research, but in integrating these traits through complementary analytical approaches that characterize physiological coordination from different perspectives. Recent studies in cotton have similarly demonstrated that integrating morphophysiological, biochemical, oxidative balance and multivariate information improves genotype discrimination under abiotic stress [16, 17, 29, 30, 49]. These cross-crop studies support the methodological value of integrating complementary trait groups, although they do not constitute direct evidence for wheat salinity responses or validate the analytical framework proposed here. Compared with conventional multivariate approaches, the cost–benefit slope analysis provides transition-specific descriptions of physiological associations, whereas IPE_ridge summarizes the combined contribution of correlated physiological modules to predefined functional responses. Accordingly, these approaches should be regarded as complementary exploratory tools that extend existing multivariate analyses rather than replace established physiological screening methods or independent validation procedures. Their greatest potential lies in supporting hypothesis generation and physiological interpretation, while future studies determine their robustness across larger and more genetically diverse wheat populations.

Several limitations should be considered when interpreting the present findings. The experiment included only three contrasting wheat genotypes evaluated under controlled lath-house conditions during a single growing season, with three biological replications and two salinity levels beyond the control. Although the salinity (EC) of the irrigation solutions was carefully adjusted and verified before application, the root-zone soil electrical conductivity, leachate electrical conductivity, drainage volume and leaching fraction were not monitored throughout the experiment. Similarly, the internal lath-house temperature and the reduction in irradiance caused by the polyethylene covering were not continuously quantified. These factors may influence the intensity and temporal dynamics of salinity stress and therefore should be considered when extrapolating the present findings to field conditions.

Another important limitation is that ion homeostasis was not directly evaluated. The maintenance of low leaf Na⁺ concentrations, K⁺ retention and a favourable K⁺:Na⁺ ratio are well-established components of salinity tolerance in wheat [23, 36]. Consequently, the present study cannot distinguish the contribution of ionic regulation from the physiological and biochemical responses investigated here. Integrating measurements of ion accumulation and compartmentation with water relations, photosynthetic performance, oxidative balance and yield will provide a more comprehensive understanding of genotype-specific salinity tolerance.

The analytical approaches proposed in this study should also be interpreted within the scope of the available dataset. The correlation analyses were based on nine (9) observations within each salinity level and no correction for multiple comparisons was applied. Accordingly, these relationships should be regarded as exploratory descriptions of association patterns rather than evidence of causal physiological networks. Similarly, the proposed cost–benefit slope analysis and IPE_ridge framework were developed and evaluated using the present dataset only. Although both approaches showed good internal consistency, they were not independently validated across additional genotype panels, environments, or growing seasons. Their broader applicability therefore remains to be established through independent validation before they can be considered robust tools for physiological screening or breeding applications.

Future research should evaluate these analytical frameworks across larger and genetically diverse wheat populations exposed to multiple salinity gradients under both controlled and field environments. Integrating physiological, biochemical and ion-homeostasis traits with genomic, phenomics and high-throughput phenotyping approaches may further improve the identification of key mechanisms underlying salinity tolerance and facilitate the development of more robust selection strategies. Recent integrated trait-based studies in cotton have similarly demonstrated that combining complementary physiological and biochemical information can improve genotype characterization under abiotic stress [16, 17, 49], although the applicability of these approaches to wheat requires further evaluation.

Taken together, the present findings indicate that salinity tolerance in the evaluated wheat genotypes was associated with the coordinated maintenance of water relations, photosynthetic function, pigment stability, compatible solute accumulation, oxidative balance and yield components rather than with any single physiological trait. Among the evaluated genotypes, BAW 1147 consistently exhibited the most favourable overall physiological performance, BARI Gom 25 showed intermediate responses and BARI Gom 28 was the most sensitive to salinity. The proposed cost–benefit slope analysis and IPE_ridge framework provide complementary exploratory approaches for integrating multiple physiological processes within the present dataset and offer additional perspectives beyond conventional multivariate analyses. However, these approaches should be regarded as hypothesis-generating and complementary analytical tools until their robustness has been confirmed through independent validation across larger wheat populations, multiple environments and direct assessment of ion homeostasis. Such validation will determine their potential utility for the physiological screening and breeding of salinity-tolerant wheat.

Conclusion

Salinity (EC) of the irrigation solution responses among the evaluated wheat genotypes were associated with coordinated changes in water relations, photosynthetic performance, pigment stability, compatible solute accumulation, oxidative balance and yield components rather than with any single physiological trait. Among the three genotypes evaluated, BAW 1147 consistently exhibited comparatively smaller physiological disruption and greater maintenance of grain yield under increasing salinity (EC) of the irrigation solution, whereas BARI Gom 25 showed intermediate responses and BARI Gom 28 was more adversely affected. The complementary cost–benefit slope analysis and IPE_ridge framework provided integrated descriptions of multitrait physiological responses within the present dataset and offered additional perspectives beyond conventional multivariate analyses. However, these approaches should be regarded as exploratory analytical tools that describe coordinated physiological responses and demonstrate internal consistency within the present dataset rather than independently validated predictive methods. Overall, the present study contributes to the physiological characterization of salinity (EC) of the irrigation solution responses in the evaluated wheat genotypes and provides a framework for generating hypotheses regarding multitrait physiological coordination under salinity stress. Future studies involving larger and more genetically diverse wheat populations, direct assessment of ion homeostasis, additional salinity environments and independent validation of the proposed analytical framework will be required before its broader applicability for physiological screening or breeding research can be established.

Supplementary Information

Supplementary Material 1.^{(700.5KB, zip)}

Acknowledgements

All the authors are grateful to the Bangladesh Agricultural Research Institute, Gazipur, Bangladesh and Bangladesh Wheat and Maize Research Institute, Dinajpur, Bangladesh, for supporting the study. The authors also wish to acknowledge Princess Nourah bint Abdulrahman University Researchers Supporting Project Number (PNURSP2026R951) Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia, for supporting the study.

Authors’ contributions

M.M.K., M.A.K., M.M.H., M.A.B.M., N.C.D.B.: conceptualization and methodology; M.M.K., H.Z., A.G., A.H.: data collection and analysis; M.M.K., M.A.K., M.M.H., M.A.B.M., N.C.D.B.: writing the manuscript; M.M.K., N.C.D.B., A.G., H.Z., A.H.: reviewing and final editing; H.Z., A.G., A.H.: supervision; and M.M.K., N.C.D.B., A.G., H.Z., A.H.: funding and project administrator. All the authors reviewed the findings and accepted the final version of the manuscript.

Funding

This research was financially supported by the Bangladesh Agricultural Research Institute, Gazipur, Bangladesh and the Bangladesh Wheat and Maize Research Institute, Dinajpur, Bangladesh. The study was also funded by Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2026R951), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia.

Data availability

The datasets generated during and/or analysed during the current study will be available from the corresponding author upon request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Md. Mustafa Khan, Email: mmkhanwrc@yahoo.com.

Akbar Hossain, Email: tanjimar2003@yahoo.com.

References

1.Adams WW III, Demmig-Adams B, Rosenstiel TN, Brightwell AK, Ebbert V. Photosynthesis and photoprotection in overwintering plants. Plant Biol (Stuttg). 2002;4(5):545–57. 10.1055/s-2002-35434. [DOI] [Google Scholar]
2.Arnon DI. Copper enzymes in isolated chloroplasts: polyphenoloxidase in Beta vulgaris. Plant Physiol. 1949;24(1):1–15. 10.1104/pp.24.1.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Ashraf M, Foolad MR. Roles of glycine betaine and proline in improving plant abiotic stress resistance. Environ Exp Bot. 2007;59(2):206–16. 10.1016/j.envexpbot.2005.12.006. [DOI] [Google Scholar]
4.Bates LS, Waldren RP, Teare ID. Rapid determination of free proline for water-stress studies. Plant Soil. 1973;39:205–7. 10.1007/BF00018060. [DOI] [Google Scholar]
5.Bisong E. Google Colaboratory. In: Building machine learning and deep learning models on Google Cloud Platform. Apress; 2019. p. 59–64. 10.1007/978-1-4842-4470-8_7. [DOI] [Google Scholar]
6.Blum A. Stress, strain, signalling and adaptation—not just a matter of definition. J Exp Bot. 2016;67(3):562–5. 10.1093/jxb/erv497. [DOI] [PubMed] [Google Scholar]
7.Blum A, Ebercon A. Cell membrane stability as a measure of drought and heat tolerance in wheat. Crop Sci. 1981;21(1):43–7. 10.2135/cropsci1981.0011183X002100010013x. [DOI] [Google Scholar]
8.Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding. Anal Biochem. 1976;72:248–54. [DOI] [PubMed] [Google Scholar]
9.Chen THH, Murata N. Glycinebetaine protects plants against abiotic stress: mechanisms and biotechnological applications. Plant Cell Environ. 2011;34(1):1–20. 10.1111/j.1365-3040.2010.02232.x. [DOI] [PubMed] [Google Scholar]
10.Csiszár J, Lantos E, Tari I, Madoşa E, Wodala B, Vashegyi Á, et al. Antioxidant enzyme activities in Allium species and their cultivars under water stress. Plant Soil Environ. 2007;53(12):517–23. [Google Scholar]
11.Dormann CF, Elith J, Bacher S, Buchmann C, Carl G, Carré G, et al. Collinearity: a review of methods to deal with it and a simulation study evaluating their performance. Ecography. 2013;36(1):27–46. 10.1111/j.1600-0587.2012.07348.x. [DOI] [Google Scholar]
12.El Sabagh A, Islam MS, Skalicky M, Ali Raza M, Singh K, Anwar Hossain M, et al. Salinity stress in wheat (Triticum aestivum L.) in the changing climate: Adaptation and management strategies. Front Agronomy. 2021;3:661932. 10.3389/fagro.2021.661932. [DOI] [Google Scholar]
13.Elia AC, Galarini R, Taticchi MI, Dörr AJM, Mantilacci L. Antioxidant responses and bioaccumulation in Ictalurus melas under mercury exposure. Ecotoxicol Environ Saf. 2003;55(2):162–7. [DOI] [PubMed] [Google Scholar]
14.Flexas J, Barbour MM, Brendel O, Cabrera HM, Carriquí M, Díaz-Espejo A, et al. Mesophyll diffusion conductance to CO₂: An unappreciated central player in photosynthesis. Plant Sci. 2012;193–194:70–84. 10.1016/j.plantsci.2012.05.009. [DOI] [PubMed] [Google Scholar]
15.Flowers TJ, Colmer TD. Plant salt tolerance: adaptations in halophytes. Ann Bot. 2015;115(3):327–31. 10.1093/aob/mcu267. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Gul N, Khan Z, Shani MY, Hafiza BS, Saeed A, Khan AI, et al. Identification of salt-resilient cotton genotypes using integrated morpho-physiological and biochemical markers at the seedling stage. Sci Rep. 2025;15:5276. 10.1038/s41598-025-89582-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Gul N, Saeed A, Khan AI, Awan FS. Dissecting seedling-stage salinity tolerance in cotton for morphophysiological and biochemical traits using multivariate analysis. SABRAO J Breed Genet. 2026;58(1):100–111. 10.54910/sabrao2026.58.1.10. [DOI]
18.Hasanuzzaman M, Bhuyan MHMB, Zulfiqar F, Raza A, Mohsin SM, Mahmud JA, et al. Reactive oxygen species and antioxidant defence in plants under abiotic stress: Revisiting the crucial role of a universal defence regulator. Antioxidants. 2020;9(8):681. 10.3390/antiox9080681. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Hoerl AE, Kennard RW. Ridge regression: biased estimation for nonorthogonal problems. Technometrics. 1970;12(1):55–67. 10.1080/00401706.1970.10488634. [DOI] [Google Scholar]
20.Ibrahimova U, Talai J, Hasan MM, Huseynova I, Raja V, Rastogi A, et al. Dissecting the osmotic and oxidative stress responses in salt-tolerant and salt-sensitive wheat genotypes under saline conditions. Plant Soil Environ. 2025;71:36–47. 10.17221/459/2024-PSE. [DOI] [Google Scholar]
21.Ibrahimova U, Suleymanova Z, Brestic M, Mammadov A, Ali OM, Abdel Latef AAH, et al. Assessing the adaptive mechanisms of two bread wheat (Triticum aestivum L.) genotypes to salinity stress. Agronomy. 2021;11:1979. 10.3390/agronomy11101979. [DOI] [Google Scholar]
22.IRRI. Statistical Tool for Agricultural Research (STAR), version 2.0.1 [Computer software]. International Rice Research Institute. 2014. https://news.irri.org/2013/08/irri-biometrics-group-releases.html
23.Ismail AM, Horie T. Genomics, physiology and molecular breeding approaches for improving salt tolerance. Annu Rev Plant Biol. 2017;68:405–34. 10.1146/annurev-arplant-042916-040936. [DOI] [PubMed] [Google Scholar]
24.Jones HG. Monitoring plant and soil water status: established and novel methods revisited and their relevance to studies of drought tolerance. J Exp Bot. 2007;58(2):119–30. 10.1093/jxb/erl118. [DOI] [PubMed] [Google Scholar]
25.Karim MR, Sultana S, Altaf-Un-Nahar M, Islam MR, Rahman F, Pretha SJ, et al. Integrated molecular defence mitigating salt stress in tomatoes using synergistic signalling molecules. Physiol Plant. 2025;177(3):e70344. 10.1111/ppl.70344. [DOI] [PubMed] [Google Scholar]
26.Kavi Kishor PB, Hima Kumari P, Sunita MSL, Sreenivasulu N. Role of proline in cell wall synthesis and plant development and its implications in plant ontogeny. Front Plant Sci. 2015;6:544. 10.3389/fpls.2015.00544. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Keunen E, Peshev D, Vangronsveld J, Van Den Ende W, Cuypers A. Plant sugars are crucial players in the oxidative challenge during abiotic stress: extending the traditional concept. Plant Cell Environ. 2013;36(7):1242–55. 10.1111/pce.12061. [DOI] [PubMed] [Google Scholar]
28.Khan MM, Karim MA, Haque MM, Mia MAB. Evaluation and selection of wheat (Triticum aestivum L.) advanced lines for salt tolerance in Bangladesh. Asian J Res Crop Sci. 2023;8(3):92–107. 10.9734/ajrcs/2023/v8i3171. [DOI] [Google Scholar]
29.Khan Z, Khan AI, Shakeel A, Awan FS. Oxidative stress physiology of cotton seedlings under drought, linking water status, pigments, osmolytes and antioxidant enzymes. Russ J Plant Physiol. 2026;73:41. 10.1134/S1021443725605701. [DOI] [Google Scholar]
30.Khan Z, Gul N, Riaz M, Bakhtawar Z, Tariq M, Khan AI, et al. Oxidative-window homeostasis in cotton under drought and heat defines climate resilience and guides genotype selection via the ROS-threshold model. Plant Growth Regul. 2026;106:5. 10.1007/s10725-025-01407-2. [DOI] [Google Scholar]
31.Lichtenthaler HK, Wellburn AR. Determinations of total carotenoids and chlorophylls a and b of leaf extracts in different solvents. Biochem Soc Trans. 1983;11(5):591–2. 10.1042/bst0110591. [DOI] [Google Scholar]
32.Microsoft Corporation. Microsoft Excel [Computer software]. 2024. https://www.microsoft.com/.
33.Mittler R. ROS are good. Trends Plant Sci. 2017;22(1):11–9. 10.1016/j.tplants.2016.08.002. [DOI] [PubMed] [Google Scholar]
34.Munns R, Gilliham M. Salinity tolerance of crops—what is the cost? New Phytol. 2015;208(3):668–73. 10.1111/nph.13519. [DOI] [PubMed] [Google Scholar]
35.Munns R, Tester M. Mechanisms of salinity tolerance. Annu Rev Plant Biol. 2008;59:651–81. 10.1146/annurev.arplant.59.032607.092911. [DOI] [PubMed] [Google Scholar]
36.Munns R, Passioura JB, Colmer TD, Byrt CS. Osmotic adjustment and energy limitations to plant growth in saline soil. New Phytol. 2020;225(3):1091–6. 10.1111/nph.15862. [DOI] [PubMed] [Google Scholar]
37.Nakano Y, Asada K. Hydrogen peroxide is scavenged by ascorbate-specific peroxidase in spinach chloroplasts. Plant Cell Physiol. 1981;22(5):867–80. 10.1093/oxfordjournals.pcp.a076232. [DOI] [Google Scholar]
38.OpenAI. ChatGPT [Large language model]. 2025. https://chatgpt.com/.
39.Patwa N, Pandey V, Gupta OP, et al. Unravelling wheat genotypic responses: insights into salinity stress tolerance in relation to oxidative stress, antioxidant mechanisms, osmolyte accumulation and grain quality parameters. BMC Plant Biol. 2024;24:875. 10.1186/s12870-024-05508-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Poorter H, Niklas KJ, Reich PB, Oleksyn J, Poot P, Mommer L. Biomass allocation to leaves, stems and roots: meta-analyses of interspecific variation and environmental control. New Phytol. 2012;193(1):30–50. 10.1111/j.1469-8137.2011.03952.x. [DOI] [PubMed] [Google Scholar]
41.Qadir M, Quillérou E, Nangia V, Murtaza G, Singh M, Thomas RJ, et al. Economics of salt-induced land degradation and restoration. Nat Resour Forum. 2014;38(4):282–95. 10.1111/1477-8947.12054. [DOI] [Google Scholar]
42.Rahman A, Uddin MN. Challenges and opportunities for saline agriculture in coastal Bangladesh. In: Future of sustainable agriculture in saline environments. CRC Press; 2021. p. 125–46. 10.1201/9781003112327-8. [DOI] [Google Scholar]
43.Raja V, Majeed U, Kang H, Rabi KI, John R. Abiotic stress: interplay between ROS, hormones and MAPKs. Environ Exp Bot. 2017;137:142–57. 10.1016/j.envexpbot.2017.02.010. [DOI] [Google Scholar]
44.Roy SJ, Negrão S, Tester M. Salt resistant crop plants. Curr Opin Biotechnol. 2014;26:115–24. 10.1016/j.copbio.2013.12.004. [DOI] [PubMed] [Google Scholar]
45.Scholander PF, Hammel HT, Bradstreet ED, Hemmingsen EA. Sap pressure in vascular plants. Science. 1965;148(3668):339–46. 10.1126/science.148.3668.339. [DOI] [PubMed] [Google Scholar]
46.Schonfeld MA, Johnson RC, Carver BF, Mornhinweg DW. Water relations in winter wheat as drought resistance indicators. Crop Sci. 1988;28(3):526–31. 10.2135/cropsci1988.0011183X002800030021x. [DOI] [Google Scholar]
47.Seleiman MF, Semida WM, Rady MM, Mohamed GF, Hemida KA, Alhammad BA, et al. Sequential application of antioxidants rectifies ion imbalance and strengthens antioxidant systems in salt-stressed cucumber. Plants (Basel). 2020;9(12):1783. 10.3390/plants9121783. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Shabala S. Learning from halophytes: physiological basis and strategies to improve abiotic stress tolerance in crops. Ann Bot. 2013;112(7):1209–21. 10.1093/aob/mct205. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Shani MY, Ashraf MY, Ramzan M, Khan Z, Batool N, Gul N, et al. Unveiling drought tolerant cotton genotypes: insights from morpho-physiological and biochemical markers at flowering. Plants (Basel). 2025;14(4):616. 10.3390/plants14040616. [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Spormann S, Nadais P, Sousa F, Pinto M, Martins M, Sousa B, et al. Accumulation of proline in plants under contaminated soils: are we on the same page? Antioxidants (Basel). 2023;12(3):666. 10.3390/antiox12030666. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Szabados L, Savouré A. Proline: a multifunctional amino acid. Trends Plant Sci. 2010;15(2):89–97. 10.1016/j.tplants.2009.11.009. [DOI] [PubMed] [Google Scholar]
52.Tardieu F. Any trait or trait-related allele can confer drought tolerance: just design the right drought scenario. J Exp Bot. 2012;63(1):25–31. 10.1093/jxb/err269. [DOI] [PubMed] [Google Scholar]
53.The jamovi project. jamovi (Version 2.5) [Computer software]. 2024. https://www.jamovi.org/.
54.Turner NC. Techniques and experimental approaches for the measurement of plant water status. Plant Soil. 1981;58:339–66. 10.1007/BF02180062. [DOI] [Google Scholar]
55.van Zelm E, Zhang Y, Testerink C. Salt tolerance mechanisms of plants. Annu Rev Plant Biol. 2020;71:403–33. 10.1146/annurev-arplant-050718-100005. [DOI] [PubMed] [Google Scholar]
56.Yang Y, Guo Y. Elucidating the molecular mechanisms mediating plant salt-stress responses. New Phytol. 2018;217(2):523–39. 10.1111/nph.14920. [DOI] [PubMed] [Google Scholar]
57.Yoshida S, Forno DA, Cock JH, Gomez KA. Laboratory manual for physiological studies of rice (3rd ed.). International Rice Research Institute, Manila, Philippines. [Book/manual. 1976. https://upa-lt.unja.ac.id/wp-content/uploads/2019/11/9711040352_content.pdf
58.Yu CW, Murphy TM, Lin CH. Hydrogen peroxide-induced chilling tolerance in mung beans mediated through ABA-independent glutathione accumulation. Funct Plant Biol. 2003;30(9):955–63. 10.1071/FP03091. [DOI] [PubMed] [Google Scholar]
59.Zulfiqar F, Akram NA, Ashraf M. Osmoprotection in plants under abiotic stresses: new insights into a classical phenomenon. Planta. 2019;251:3. 10.1007/s00425-019-03293-1. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1.^{(700.5KB, zip)}

Data Availability Statement

The datasets generated during and/or analysed during the current study will be available from the corresponding author upon request.

[CR1] 1.Adams WW III, Demmig-Adams B, Rosenstiel TN, Brightwell AK, Ebbert V. Photosynthesis and photoprotection in overwintering plants. Plant Biol (Stuttg). 2002;4(5):545–57. 10.1055/s-2002-35434. [DOI] [Google Scholar]

[CR2] 2.Arnon DI. Copper enzymes in isolated chloroplasts: polyphenoloxidase in Beta vulgaris. Plant Physiol. 1949;24(1):1–15. 10.1104/pp.24.1.1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Ashraf M, Foolad MR. Roles of glycine betaine and proline in improving plant abiotic stress resistance. Environ Exp Bot. 2007;59(2):206–16. 10.1016/j.envexpbot.2005.12.006. [DOI] [Google Scholar]

[CR4] 4.Bates LS, Waldren RP, Teare ID. Rapid determination of free proline for water-stress studies. Plant Soil. 1973;39:205–7. 10.1007/BF00018060. [DOI] [Google Scholar]

[CR5] 5.Bisong E. Google Colaboratory. In: Building machine learning and deep learning models on Google Cloud Platform. Apress; 2019. p. 59–64. 10.1007/978-1-4842-4470-8_7. [DOI] [Google Scholar]

[CR6] 6.Blum A. Stress, strain, signalling and adaptation—not just a matter of definition. J Exp Bot. 2016;67(3):562–5. 10.1093/jxb/erv497. [DOI] [PubMed] [Google Scholar]

[CR7] 7.Blum A, Ebercon A. Cell membrane stability as a measure of drought and heat tolerance in wheat. Crop Sci. 1981;21(1):43–7. 10.2135/cropsci1981.0011183X002100010013x. [DOI] [Google Scholar]

[CR8] 8.Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding. Anal Biochem. 1976;72:248–54. [DOI] [PubMed] [Google Scholar]

[CR9] 9.Chen THH, Murata N. Glycinebetaine protects plants against abiotic stress: mechanisms and biotechnological applications. Plant Cell Environ. 2011;34(1):1–20. 10.1111/j.1365-3040.2010.02232.x. [DOI] [PubMed] [Google Scholar]

[CR10] 10.Csiszár J, Lantos E, Tari I, Madoşa E, Wodala B, Vashegyi Á, et al. Antioxidant enzyme activities in Allium species and their cultivars under water stress. Plant Soil Environ. 2007;53(12):517–23. [Google Scholar]

[CR11] 11.Dormann CF, Elith J, Bacher S, Buchmann C, Carl G, Carré G, et al. Collinearity: a review of methods to deal with it and a simulation study evaluating their performance. Ecography. 2013;36(1):27–46. 10.1111/j.1600-0587.2012.07348.x. [DOI] [Google Scholar]

[CR12] 12.El Sabagh A, Islam MS, Skalicky M, Ali Raza M, Singh K, Anwar Hossain M, et al. Salinity stress in wheat (Triticum aestivum L.) in the changing climate: Adaptation and management strategies. Front Agronomy. 2021;3:661932. 10.3389/fagro.2021.661932. [DOI] [Google Scholar]

[CR13] 13.Elia AC, Galarini R, Taticchi MI, Dörr AJM, Mantilacci L. Antioxidant responses and bioaccumulation in Ictalurus melas under mercury exposure. Ecotoxicol Environ Saf. 2003;55(2):162–7. [DOI] [PubMed] [Google Scholar]

[CR14] 14.Flexas J, Barbour MM, Brendel O, Cabrera HM, Carriquí M, Díaz-Espejo A, et al. Mesophyll diffusion conductance to CO₂: An unappreciated central player in photosynthesis. Plant Sci. 2012;193–194:70–84. 10.1016/j.plantsci.2012.05.009. [DOI] [PubMed] [Google Scholar]

[CR15] 15.Flowers TJ, Colmer TD. Plant salt tolerance: adaptations in halophytes. Ann Bot. 2015;115(3):327–31. 10.1093/aob/mcu267. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.Gul N, Khan Z, Shani MY, Hafiza BS, Saeed A, Khan AI, et al. Identification of salt-resilient cotton genotypes using integrated morpho-physiological and biochemical markers at the seedling stage. Sci Rep. 2025;15:5276. 10.1038/s41598-025-89582-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Gul N, Saeed A, Khan AI, Awan FS. Dissecting seedling-stage salinity tolerance in cotton for morphophysiological and biochemical traits using multivariate analysis. SABRAO J Breed Genet. 2026;58(1):100–111. 10.54910/sabrao2026.58.1.10. [DOI]

[CR18] 18.Hasanuzzaman M, Bhuyan MHMB, Zulfiqar F, Raza A, Mohsin SM, Mahmud JA, et al. Reactive oxygen species and antioxidant defence in plants under abiotic stress: Revisiting the crucial role of a universal defence regulator. Antioxidants. 2020;9(8):681. 10.3390/antiox9080681. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Hoerl AE, Kennard RW. Ridge regression: biased estimation for nonorthogonal problems. Technometrics. 1970;12(1):55–67. 10.1080/00401706.1970.10488634. [DOI] [Google Scholar]

[CR20] 20.Ibrahimova U, Talai J, Hasan MM, Huseynova I, Raja V, Rastogi A, et al. Dissecting the osmotic and oxidative stress responses in salt-tolerant and salt-sensitive wheat genotypes under saline conditions. Plant Soil Environ. 2025;71:36–47. 10.17221/459/2024-PSE. [DOI] [Google Scholar]

[CR21] 21.Ibrahimova U, Suleymanova Z, Brestic M, Mammadov A, Ali OM, Abdel Latef AAH, et al. Assessing the adaptive mechanisms of two bread wheat (Triticum aestivum L.) genotypes to salinity stress. Agronomy. 2021;11:1979. 10.3390/agronomy11101979. [DOI] [Google Scholar]

[CR22] 22.IRRI. Statistical Tool for Agricultural Research (STAR), version 2.0.1 [Computer software]. International Rice Research Institute. 2014. https://news.irri.org/2013/08/irri-biometrics-group-releases.html

[CR23] 23.Ismail AM, Horie T. Genomics, physiology and molecular breeding approaches for improving salt tolerance. Annu Rev Plant Biol. 2017;68:405–34. 10.1146/annurev-arplant-042916-040936. [DOI] [PubMed] [Google Scholar]

[CR24] 24.Jones HG. Monitoring plant and soil water status: established and novel methods revisited and their relevance to studies of drought tolerance. J Exp Bot. 2007;58(2):119–30. 10.1093/jxb/erl118. [DOI] [PubMed] [Google Scholar]

[CR25] 25.Karim MR, Sultana S, Altaf-Un-Nahar M, Islam MR, Rahman F, Pretha SJ, et al. Integrated molecular defence mitigating salt stress in tomatoes using synergistic signalling molecules. Physiol Plant. 2025;177(3):e70344. 10.1111/ppl.70344. [DOI] [PubMed] [Google Scholar]

[CR26] 26.Kavi Kishor PB, Hima Kumari P, Sunita MSL, Sreenivasulu N. Role of proline in cell wall synthesis and plant development and its implications in plant ontogeny. Front Plant Sci. 2015;6:544. 10.3389/fpls.2015.00544. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR27] 27.Keunen E, Peshev D, Vangronsveld J, Van Den Ende W, Cuypers A. Plant sugars are crucial players in the oxidative challenge during abiotic stress: extending the traditional concept. Plant Cell Environ. 2013;36(7):1242–55. 10.1111/pce.12061. [DOI] [PubMed] [Google Scholar]

[CR28] 28.Khan MM, Karim MA, Haque MM, Mia MAB. Evaluation and selection of wheat (Triticum aestivum L.) advanced lines for salt tolerance in Bangladesh. Asian J Res Crop Sci. 2023;8(3):92–107. 10.9734/ajrcs/2023/v8i3171. [DOI] [Google Scholar]

[CR29] 29.Khan Z, Khan AI, Shakeel A, Awan FS. Oxidative stress physiology of cotton seedlings under drought, linking water status, pigments, osmolytes and antioxidant enzymes. Russ J Plant Physiol. 2026;73:41. 10.1134/S1021443725605701. [DOI] [Google Scholar]

[CR30] 30.Khan Z, Gul N, Riaz M, Bakhtawar Z, Tariq M, Khan AI, et al. Oxidative-window homeostasis in cotton under drought and heat defines climate resilience and guides genotype selection via the ROS-threshold model. Plant Growth Regul. 2026;106:5. 10.1007/s10725-025-01407-2. [DOI] [Google Scholar]

[CR31] 31.Lichtenthaler HK, Wellburn AR. Determinations of total carotenoids and chlorophylls a and b of leaf extracts in different solvents. Biochem Soc Trans. 1983;11(5):591–2. 10.1042/bst0110591. [DOI] [Google Scholar]

[CR32] 32.Microsoft Corporation. Microsoft Excel [Computer software]. 2024. https://www.microsoft.com/.

[CR33] 33.Mittler R. ROS are good. Trends Plant Sci. 2017;22(1):11–9. 10.1016/j.tplants.2016.08.002. [DOI] [PubMed] [Google Scholar]

[CR34] 34.Munns R, Gilliham M. Salinity tolerance of crops—what is the cost? New Phytol. 2015;208(3):668–73. 10.1111/nph.13519. [DOI] [PubMed] [Google Scholar]

[CR35] 35.Munns R, Tester M. Mechanisms of salinity tolerance. Annu Rev Plant Biol. 2008;59:651–81. 10.1146/annurev.arplant.59.032607.092911. [DOI] [PubMed] [Google Scholar]

[CR36] 36.Munns R, Passioura JB, Colmer TD, Byrt CS. Osmotic adjustment and energy limitations to plant growth in saline soil. New Phytol. 2020;225(3):1091–6. 10.1111/nph.15862. [DOI] [PubMed] [Google Scholar]

[CR37] 37.Nakano Y, Asada K. Hydrogen peroxide is scavenged by ascorbate-specific peroxidase in spinach chloroplasts. Plant Cell Physiol. 1981;22(5):867–80. 10.1093/oxfordjournals.pcp.a076232. [DOI] [Google Scholar]

[CR38] 38.OpenAI. ChatGPT [Large language model]. 2025. https://chatgpt.com/.

[CR39] 39.Patwa N, Pandey V, Gupta OP, et al. Unravelling wheat genotypic responses: insights into salinity stress tolerance in relation to oxidative stress, antioxidant mechanisms, osmolyte accumulation and grain quality parameters. BMC Plant Biol. 2024;24:875. 10.1186/s12870-024-05508-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] 40.Poorter H, Niklas KJ, Reich PB, Oleksyn J, Poot P, Mommer L. Biomass allocation to leaves, stems and roots: meta-analyses of interspecific variation and environmental control. New Phytol. 2012;193(1):30–50. 10.1111/j.1469-8137.2011.03952.x. [DOI] [PubMed] [Google Scholar]

[CR41] 41.Qadir M, Quillérou E, Nangia V, Murtaza G, Singh M, Thomas RJ, et al. Economics of salt-induced land degradation and restoration. Nat Resour Forum. 2014;38(4):282–95. 10.1111/1477-8947.12054. [DOI] [Google Scholar]

[CR42] 42.Rahman A, Uddin MN. Challenges and opportunities for saline agriculture in coastal Bangladesh. In: Future of sustainable agriculture in saline environments. CRC Press; 2021. p. 125–46. 10.1201/9781003112327-8. [DOI] [Google Scholar]

[CR43] 43.Raja V, Majeed U, Kang H, Rabi KI, John R. Abiotic stress: interplay between ROS, hormones and MAPKs. Environ Exp Bot. 2017;137:142–57. 10.1016/j.envexpbot.2017.02.010. [DOI] [Google Scholar]

[CR44] 44.Roy SJ, Negrão S, Tester M. Salt resistant crop plants. Curr Opin Biotechnol. 2014;26:115–24. 10.1016/j.copbio.2013.12.004. [DOI] [PubMed] [Google Scholar]

[CR45] 45.Scholander PF, Hammel HT, Bradstreet ED, Hemmingsen EA. Sap pressure in vascular plants. Science. 1965;148(3668):339–46. 10.1126/science.148.3668.339. [DOI] [PubMed] [Google Scholar]

[CR46] 46.Schonfeld MA, Johnson RC, Carver BF, Mornhinweg DW. Water relations in winter wheat as drought resistance indicators. Crop Sci. 1988;28(3):526–31. 10.2135/cropsci1988.0011183X002800030021x. [DOI] [Google Scholar]

[CR47] 47.Seleiman MF, Semida WM, Rady MM, Mohamed GF, Hemida KA, Alhammad BA, et al. Sequential application of antioxidants rectifies ion imbalance and strengthens antioxidant systems in salt-stressed cucumber. Plants (Basel). 2020;9(12):1783. 10.3390/plants9121783. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Shabala S. Learning from halophytes: physiological basis and strategies to improve abiotic stress tolerance in crops. Ann Bot. 2013;112(7):1209–21. 10.1093/aob/mct205. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Shani MY, Ashraf MY, Ramzan M, Khan Z, Batool N, Gul N, et al. Unveiling drought tolerant cotton genotypes: insights from morpho-physiological and biochemical markers at flowering. Plants (Basel). 2025;14(4):616. 10.3390/plants14040616. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR50] 50.Spormann S, Nadais P, Sousa F, Pinto M, Martins M, Sousa B, et al. Accumulation of proline in plants under contaminated soils: are we on the same page? Antioxidants (Basel). 2023;12(3):666. 10.3390/antiox12030666. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] 51.Szabados L, Savouré A. Proline: a multifunctional amino acid. Trends Plant Sci. 2010;15(2):89–97. 10.1016/j.tplants.2009.11.009. [DOI] [PubMed] [Google Scholar]

[CR52] 52.Tardieu F. Any trait or trait-related allele can confer drought tolerance: just design the right drought scenario. J Exp Bot. 2012;63(1):25–31. 10.1093/jxb/err269. [DOI] [PubMed] [Google Scholar]

[CR53] 53.The jamovi project. jamovi (Version 2.5) [Computer software]. 2024. https://www.jamovi.org/.

[CR54] 54.Turner NC. Techniques and experimental approaches for the measurement of plant water status. Plant Soil. 1981;58:339–66. 10.1007/BF02180062. [DOI] [Google Scholar]

[CR55] 55.van Zelm E, Zhang Y, Testerink C. Salt tolerance mechanisms of plants. Annu Rev Plant Biol. 2020;71:403–33. 10.1146/annurev-arplant-050718-100005. [DOI] [PubMed] [Google Scholar]

[CR56] 56.Yang Y, Guo Y. Elucidating the molecular mechanisms mediating plant salt-stress responses. New Phytol. 2018;217(2):523–39. 10.1111/nph.14920. [DOI] [PubMed] [Google Scholar]

[CR57] 57.Yoshida S, Forno DA, Cock JH, Gomez KA. Laboratory manual for physiological studies of rice (3rd ed.). International Rice Research Institute, Manila, Philippines. [Book/manual. 1976. https://upa-lt.unja.ac.id/wp-content/uploads/2019/11/9711040352_content.pdf

[CR58] 58.Yu CW, Murphy TM, Lin CH. Hydrogen peroxide-induced chilling tolerance in mung beans mediated through ABA-independent glutathione accumulation. Funct Plant Biol. 2003;30(9):955–63. 10.1071/FP03091. [DOI] [PubMed] [Google Scholar]

[CR59] 59.Zulfiqar F, Akram NA, Ashraf M. Osmoprotection in plants under abiotic stresses: new insights into a classical phenomenon. Planta. 2019;251:3. 10.1007/s00425-019-03293-1. [DOI] [PubMed] [Google Scholar]

PERMALINK

Integrating multitrait physiological approaches to characterize salinity tolerance in bread wheat (Triticum aestivum L.) genotypes

Md Mustafa Khan

Md Abdul Karim

Md Moynul Haque

M A Baset Mia

Naresh Chandra Deb Barma

Hela Znazen

Ahmed Gaber

Akbar Hossain

Abstract

Background

Results

Conclusions

Supplementary Information

Introduction

Materials and methods

Experimental site

Plant materials and salinity treatment application

Experimental design and seed sowing

Fertilizer application

Intercultural operations

Plant water relations

Measurement of leaf water potential

Measurement of relative water content

Compatible solute estimation

Estimation of proline content

Estimation of soluble sugars

Gas exchange measurements

Photosynthetic pigment determination

Oxidative stress indicators

Estimation of H2O2

Estimation of cell membrane injury

Determination of antioxidant enzyme activities

Catalase (CAT)

Ascorbate peroxidase (APX)

Glutathione peroxidase (GPX)

Yield and yield attributes

Statistical analysis

Integrated physiological efficiency (IPEridge) framework

Standardization of physiological traits

Cost–benefit slope analysis (ΔBenefit/ΔCost)

Integrated physiological efficiency (IPEridge)

Results

Plant water relations (LWP and RWC)

Fig. 1.

Compatible solutes

Fig. 2.

Leaf gas exchange

Fig. 3.

Photosynthetic pigments

Fig. 4.

Oxidative stress responses (H2O2 and cell membrane injury)

Fig. 5.

Antioxidant enzyme activities (CAT, APX and GPX)

Fig. 6.

Grain yield and yield components

Fig. 7.

Table 1.

Correlation dynamics of grain yield under salinity stress

Fig. 8.

Fig. 9.

Cost–benefit relationships and integrated physiological efficiency

Cost–benefit slope analysis for Pn and GY

Fig. 10.

Integrated physiological efficiency (IPEridge) and its association with grain yield and photosynthetic performance

Fig. 11.

Discussion

Conclusion

Supplementary Information

Acknowledgements

Authors’ contributions

Funding

Data availability

Declarations

Ethics approval and consent to participate

Consent for publication

Competing interests

Footnotes

Contributor Information

Estimation of H₂O₂

Integrated physiological efficiency (IPE_ridge) framework

Integrated physiological efficiency (IPE_ridge)

Oxidative stress responses (H₂O₂ and cell membrane injury)

Cost–benefit slope analysis for P_n and GY

Integrated physiological efficiency (IPE_ridge) and its association with grain yield and photosynthetic performance